Skip to content

REST API

The web app uses the SREInsights REST API for everything it does, and you can use the same API from scripts, CI pipelines and other tools. This page lists every endpoint under /api/v1, grouped by area, with the request bodies and error codes that matter.

Runners use a separate API, /runner/v1. It is described in Runner protocol.

https://console.sreinsights.com/api/v1

On a self-hosted instance, use <your SREInsights URL>/api/v1.

Every request and response body is JSON. Send content-type: application/json with a body. A request body may be at most 1 MB.

A successful call answers with status: "success" and the result in data:

{ "status": "success", "data": { "id": "0193…", "name": "Checkout API" } }

Calls that create something answer 201. Calls that queue work for later (Run now, Backfill, a report send, a webhook trigger) answer 202. A delete answers 200 with "data": null.

Every response carries an x-request-id header. Send your own x-request-id (8–64 letters, digits, _ or -) to correlate a call with your logs; otherwise SREInsights makes one.

Most endpoints need a signed-in user and an active organization. You sign in with an email and password, and get back two tokens:

Token Where it is Lifetime What it is for
Access token data.accessToken in the response body 15 minutes by default (SREINSIGHTS_ACCESS_TOKEN_TTL_S) Send it on every call as Authorization: Bearer <access token>.
Refresh token An HttpOnly, SameSite=Strict cookie named sreinsights_rt, scoped to /api/v1/auth 30 days by default (SREINSIGHTS_REFRESH_TOKEN_TTL_DAYS) Trade it for a new access token with POST /auth/refresh.

The cookie is marked Secure when the instance’s URL starts with https://. The lifetimes above are set by the operator; see Environment variables.

POST /auth/login with { "email": "...", "password": "..." }. A wrong email or password, or a deactivated account, all answer 401 INVALID_CREDENTIALS.

When a call answers 401 TOKEN_EXPIRED, call POST /auth/refresh with the refresh cookie. You get a new access token and a new refresh cookie: each refresh token works once. A refresh token that is unknown, revoked or expired answers 401 REFRESH_INVALID, and the cookie is cleared. Sign in again.

An access token carries one active organization. Sign-in picks the first organization you joined. To work in another one:

  1. List your organizations with GET /auth/me (field orgs) or GET /orgs/summary.
  2. Call POST /orgs/switch with { "orgId": "<org id>" }.
  3. Use the accessToken it returns from then on.

An organization you are not a member of answers 404 NOT_FOUND. A refresh keeps the organization you switched to, as long as you are still a member.

Endpoints that read or write an organization’s data answer 409 NO_ACTIVE_ORG when the token has no active organization (for example, you were removed from your last one).

Status Code What it means
401 UNAUTHENTICATED No Authorization: Bearer header.
401 TOKEN_EXPIRED The access token expired. Refresh it.
401 TOKEN_INVALID The token is malformed or not signed by this instance. Sign in again.
401 TOKEN_REVOKED The session ended: you signed out, changed your password, or your account was deactivated. Sign in again.
403 ORG_ACCESS_REVOKED You are no longer a member of the token’s organization, or it was deleted. Switch organizations.
409 NO_ACTIVE_ORG The endpoint needs an organization and the token has none.

Sessions are checked on every call, so a sign-out or password change takes effect at once, not when the access token expires.

Three endpoints don’t take a user’s token. Each uses a token of its own:

Endpoint Token
POST /hooks/{triggerId} The webhook trigger’s pwt_… token. See Webhooks.
POST /ingest/{type}/{connectionId} The connection’s pit_… ingest token. See Ingest.
GET /public/status/{token} None: the token in the address is the key. See Status pages.

Every error has the same shape:

{
"status": "fail",
"code": "VALIDATION_FAILED",
"message": "Some fields are missing or invalid",
"details": [{ "path": "schedule.cron", "message": "Required" }],
"requestId": "4f1c2b8e-…"
}
Field What it is
status fail for a 4xx answer, error for a 5xx answer.
code A stable machine-readable code. Branch on this, not on message.
message A sentence you can show to a person.
details Present on some errors. For VALIDATION_FAILED, a list of { path, message }, one per field.
requestId The same value as the x-request-id header. Quote it when you report a problem.

These codes can come from any endpoint:

Status Code What it means
400 VALIDATION_FAILED A field is missing or invalid. details names each one.
400 BAD_REQUEST The body is not valid JSON.
403 POLICY_DENIED Your role doesn’t allow this action. See Organizations and roles.
404 NOT_FOUND No such record in your organization. A record in another organization or client also answers 404, never 403.
404 ROUTE_NOT_FOUND There is no endpoint with this method and path.
409 ALREADY_EXISTS A record with these values already exists, for example a runner group with the same name.
409 IN_USE Other records still depend on this one. Move or delete them first.
413 PAYLOAD_TOO_LARGE The request body is larger than 1 MB.
500 INTERNAL Something failed on the server. Try again, and quote the requestId if it keeps happening.

Each area below lists the codes its endpoints add.

The API has no page cursors. Lists return everything up to a fixed cap: monitors up to 1,000, silences up to 500.

Three history endpoints take a limit and, where shown, a before time to page back:

Endpoint limit before
GET /monitors/{id}/results 1–200, default 50 ISO date-time; results that finished before it
GET /jobs/{id}/runs 1–200, default 50 ISO date-time; runs created before it
GET /alerts 1–200, default 50 Not supported

To page back, pass the oldest finishedAt (results) or createdAt (runs) you got as before.

Only the public status page endpoint is rate limited: 60 requests per minute per client address, and 1,200 per minute in all, counted per server process. Over the limit, it answers 429 with a plain-text body and a retry-after header in seconds.

The AI endpoints pass on the AI service’s own limit as 429 AI_RATE_LIMITED.

Method Path What it does
POST /auth/signup Public. Creates an account and a new organization with you as Owner. Answers 201 with an access token and the refresh cookie.
POST /auth/login Public. Signs in. Answers with an access token and the refresh cookie.
POST /auth/refresh Public (needs the refresh cookie). A new access token and refresh cookie.
POST /auth/logout Ends this session and clears the refresh cookie.
GET /auth/me You, your organizations (with your role in each), the active organization, its plan, and which pages your role may see.
POST /auth/password Changes your password. Your other sessions are signed out; this one gets a new access token and cookie.
POST /auth/delete-account Deletes your account for good.
GET /auth/reset?token=… Public. Who a password reset link is for.
POST /auth/reset Public. Sets a new password from a reset link and signs you in.
GET /auth/invite?token=… Public. Who an invitation is for.
POST /auth/accept-invite Public, or signed in as the invited email. Joins the organization and answers with a session in it.

Request bodies:

Endpoint Body
POST /auth/signup { "email", "password", "name", "orgName" }
POST /auth/login { "email", "password" }
POST /auth/password { "currentPassword", "newPassword" }
POST /auth/delete-account { "password" }
POST /auth/reset { "token", "password" }
POST /auth/accept-invite { "token", "name"?, "password"? }. name and password are needed when the invited email has no account yet. Answers 201 when it created the account, 200 otherwise.

Passwords must be at least 10 characters.

Status Code When
400 PASSWORD_TOO_SHORT A new password is under 10 characters.
400 INVALID_CREDENTIALS The current password is wrong (password change, account deletion).
400 PASSWORD_UNCHANGED The new password is the same as the current one.
400 ACCOUNT_DETAILS_REQUIRED Accepting an invitation without an account, and name or password is missing.
401 INVALID_CREDENTIALS Sign-in failed.
404 INVITE_INVALID, RESET_INVALID The invitation or reset link was used, revoked or never existed.
409 EMAIL_TAKEN An account with this email already exists.
409 SIGN_IN_TO_ACCEPT The invited email has an account: sign in as it, then accept.
409 LAST_OWNER You are the only Owner of an organization. Make someone else Owner first.
410 INVITE_EXPIRED, RESET_EXPIRED The link expired. Ask for a new one.
Method Path What it does
GET /orgs/summary Every organization you belong to, with its health counts. Needs a session, not an active organization.
POST /orgs/switch Makes another organization active. Body { "orgId" }. Answers { "accessToken" }.
GET /members The active organization’s members and open invitations.
GET /roles Every role, each with assignable: whether your role may give it.
POST /members Adds a member by email. Body { "email", "role" }.
PATCH /members/{userId} Changes a member’s role, or the clients they are limited to. Body { "role"?, "clientIds"? }.
POST /members/{userId}/reset-password Makes a one-time password reset link for a member, valid for 24 hours. Shown once.
DELETE /members/{userId} Removes a member from the organization.
DELETE /invites/{id} Revokes an open invitation.

POST /members answers 201 in one of two ways:

  • The email already has an account: they join at once, and data.joined is true.
  • It doesn’t: data.joined is false and data.link is a one-time /join?token=… link, valid for 7 days. Send it to them yourself.

In PATCH /members/{userId}, clientIds: [] means every client. Owners and Admins always see every client. See Clients.

Status Code When
400 UNKNOWN_ROLE No role with that name.
400 CLIENT_LIMIT_NOT_APPLICABLE You tried to limit an Owner or Admin to some clients.
400 UNKNOWN_CLIENT One of clientIds is not in this organization.
403 POLICY_DENIED Your role can’t manage members, or can’t grant or change a role above your own.
409 ALREADY_MEMBER The email is already a member.
409 LAST_OWNER The change would leave the organization without an Owner.
Method Path What it does
GET /runner-group-kinds Which kinds of runner group this instance offers: polling, and lambda when the operator switched Lambda runner groups on.
GET /runner-groups Your runner groups, each with its runners and their status.
GET /hosted-runner-groups Hosted runner locations every organization may use. Read-only: the operator runs them.
POST /runner-groups Creates a runner group.
PATCH /runner-groups/{id} Changes the description, or a Lambda group’s function.
DELETE /runner-groups/{id} Deletes a group no monitor runs on.
POST /runner-groups/{id}/registration-tokens Makes a one-time registration token (prt_…) and the install command. The token is shown once.
POST /runners/{id}/disable Disables a runner. Its running work goes back to the group at once. Answers { "disabled": true, "requeuedRuns" }.

POST /runner-groups body:

Field Type Required Default What it does
name string Yes Lower-case letters, digits, ., _ or -, up to 63 characters.
description string No Up to 500 characters.
kind polling or lambda No polling Fixed once the group exists.
lambda object For lambda { "functionName", "region" }. functionName is a name, name:alias or a function ARN.

POST /runner-groups/{id}/registration-tokens body (optional):

Field Type Default What it does
uses integer, 1–50 1 How many runners may register with the token.
expiresInHours integer, 1–168 24 When the token stops working.

The response holds token, group, usesLeft, expiresAt and install. For a polling group, install.docker is a ready docker run command. For a Lambda group, install.lambda.identityCommand registers the function once and prints its identity; see Set up a Lambda runner.

Each runner in GET /runner-groups has a status: online (it called in during the last 60 seconds), offline, disabled, or on_demand for a Lambda group’s function. A group’s redundant is true when two or more of its runners are online.

Status Code When
400 LAMBDA_RUNNERS_DISABLED Lambda runner groups aren’t switched on for this instance.
400 LAMBDA_ROLE_NOT_SUPPORTED You sent roleArn or externalId. Invoking a function through a role in another AWS account isn’t available.
400 NOT_A_LAMBDA_GROUP You sent lambda for a polling group.
409 IN_USE Monitors still run on the group. Move them first.
409 ALREADY_EXISTS A group with this name exists.
Method Path What it does
GET /monitors All monitors, by name.
POST /monitors Creates a monitor and its schedule.
GET /monitors/{id} One monitor.
PATCH /monitors/{id} Changes any of the fields below. Send only what changes.
DELETE /monitors/{id} Deletes a monitor. Its schedule and webhook triggers stop at once.
POST /monitors/{id}/run Run now: queues one run, even while the monitor is paused. Answers 202.
GET /monitors/{id}/results Results, newest first. Takes limit and before.

POST /monitors body:

Field Type Required Default What it does
name string Yes Up to 120 characters.
description string No Up to 2,000 characters.
clientId uuid No The client it belongs to. Leave out for the organization’s own.
enabled boolean No true false pauses the schedule.
definition object Yes What to check. See Check types and Flow steps.
parameters array No [] Inputs each run receives as params.<name>. See Monitors.
runsOn object Yes { "group": "<group name or id>", "labels": [] }. A hosted location’s name works too.
schedule object Yes { "cron", "timezone": "UTC", "inputs": {} }. inputs are parameter values for scheduled runs.
labels object No {} Up to 30 name: value pairs. Names are lower-case letters, digits and _.

The cron expression takes 5 fields, 6 with seconds, or a macro such as @hourly. The timezone is an IANA name such as Europe/Berlin.

PATCH /monitors/{id} with "enabled": false also resolves the monitor’s open alerts. Leaving schedule.inputs out of a PATCH keeps the current values.

POST /monitors/{id}/run body: { "params": { … } } (optional). The run uses the schedule’s inputs, overridden by params. It answers { "jobId", "status": "queued", "queueExpiresAt", "params" }. A queued run waits up to 5 minutes for a runner.

A monitor’s activeRun is null, or { "state": "waiting_for_runner" | "running", "since" }.

Status Code When
400 INVALID_SCHEDULE The cron expression or timezone doesn’t parse, or the schedule never runs.
400 SCHEDULE_TOO_FREQUENT The schedule runs more often than every 30 seconds, or more often than this instance allows.
400 UNKNOWN_RUNNER_GROUP runsOn.group names no group of yours and no hosted location.
400 UNKNOWN_CLIENT clientId is not in this organization.
400 JOB_TYPE_ON_MONITOR The definition is a data job. Create it with POST /jobs.
400 HOSTED_SCRIPT_NOT_ALLOWED The monitor runs on a hosted location and uses a script.
400 HOSTED_RUNNER_LOCAL_SECRETS A connection keeps its secrets on your runners, so hosted runners can’t use it.
400 LAMBDA_DOCKER_SCRIPT_NOT_ALLOWED A script runs in a Docker container on a Lambda runner group. Set its runtime to local.
400 VALIDATION_FAILED Also when a connection the definition names doesn’t exist, belongs to another client, or is the wrong type for its step.
403 PLAN_LIMIT On hosted runners: the schedule is more frequent than your plan allows, or you are at your plan’s number of hosted monitors.

A webhook trigger is a URL that starts a run of its monitor, for example from a deploy pipeline. See Trigger a monitor from a webhook.

Method Path What it does
GET /monitors/{id}/triggers The monitor’s webhook triggers.
POST /monitors/{id}/triggers Creates one. Body { "name", "enabled": true, "inputs": {} }. Answers with the url and a token (pwt_…), shown once.
PATCH /monitors/{id}/triggers/{triggerId} Changes name, enabled or inputs.
POST /monitors/{id}/triggers/{triggerId}/token Makes a new token. The old one stops working. The new one is shown once.
DELETE /monitors/{id}/triggers/{triggerId} Deletes a trigger.
POST /hooks/{triggerId} Calls the trigger. No user sign-in: authenticated by the trigger’s token. Answers 202.

A monitor can have at most 10 webhook triggers (400 TOO_MANY_TRIGGERS).

Calling POST /hooks/{triggerId}:

  • Send the token as Authorization: Bearer pwt_…, or as X-SREInsights-Token: pwt_… when the sender can’t set Authorization.
  • The JSON body, up to 64 KB, reaches the check as trigger.body.
  • Parameter values come from the trigger’s inputs, overridden by the body’s params object. A sender that can’t nest may send them as top-level fields named like the parameters.
  • One run per monitor at a time. If a run is already queued or running, the answer is { "queued": false, "reason": "…" }. Otherwise it is { "queued": true, "jobId", "params" }.
Status Code When
400 VALIDATION_FAILED A parameter value is missing or invalid. details names it under params.
401 TRIGGER_UNAUTHORIZED The token is missing or wrong, or the trigger or its monitor was deleted.
409 TRIGGER_DISABLED The trigger is turned off.
409 MONITOR_PAUSED The monitor is paused.
413 TRIGGER_BODY_TOO_LARGE The body is larger than 64 KB.

Alert rules decide when a monitor’s results raise an alert. Field reference: Alert rules.

Method Path What it does
GET /monitors/{id}/rules The monitor’s rules, each with its last evaluation.
POST /monitors/{id}/rules Creates a rule.
POST /monitors/{id}/rules/preview Tries a condition against the monitor’s last 3,000 results, without saving.
PATCH /alert-rules/{id} Changes a rule.
DELETE /alert-rules/{id} Deletes a rule.

POST /monitors/{id}/rules body:

Field Type Required Default What it does
name string Yes Up to 120 characters.
enabled boolean No true
severity string Yes See Severity tiers.
condition object Yes When the rule matches.
resolveAfter integer, 1–20 No 2 Evaluations in a row that must not match before the alert resolves.
reopenWindowSeconds integer, 0–86400 No 1800 An alert that fires again within this time reopens instead of starting over.
labels object No {} Copied onto the alert.
annotations object No summary, description, runbook_url.

The preview body is { "condition", "resolveAfter" }. A condition whose expression doesn’t parse answers 400 INVALID_EXPRESSION, with the field’s path in details.

Method Path What it does
GET /alerts Alerts, filtered by the query below.
GET /alerts/{id} One alert with its events and notification deliveries.
POST /alerts/{id}/acknowledge Acknowledges a firing alert. Body { "note"? }.
POST /alerts/{id}/resolve Resolves an open alert. Body { "note"? }.

GET /alerts query:

Parameter Values Default
state open, resolved, all open (firing or acknowledged)
severity a severity
monitorId uuid
clientId uuid, or none for the organization’s own
limit 1–200 50

A note is up to 1,000 characters.

Status Code When
409 ALERT_NOT_FIRING Only a firing alert can be acknowledged.
409 ALERT_RESOLVED The alert is already resolved.

A silence holds back notifications for matching alerts for a while. See Silences.

Method Path What it does
GET /silences Silences. ?state=current (default: active or upcoming), active, upcoming, expired or all.
POST /silences Creates a silence.
POST /silences/{id}/expire Ends a silence now.

POST /silences body:

Field Type Required What it does
matchers object Yes Any of severity, source, monitorIds, clientIds, labels. At least one.
startsAt ISO date-time with offset No Defaults to now.
endsAt ISO date-time with offset One of these two When it ends.
durationMinutes integer One of these two How long it lasts.
comment string Yes Why. Up to 1,000 characters.

A silence lasts at most 30 days. One that ends before it starts, or is already over, answers 400 INVALID_SILENCE.

Channels say where a notice goes. The policy is the ordered list of routes that decides which channels get which alerts. See Notifications.

Method Path What it does
GET /notification-channels Channels.
POST /notification-channels Creates a channel. Body { "name", "connectionId", "params": {}, "enabled": true }.
PATCH /notification-channels/{id} Changes name, enabled, or params (replaced as a whole).
DELETE /notification-channels/{id} Deletes a channel. Routes that name it stop delivering to it.
GET /notification-policy The routes, with version and updatedAt.
PUT /notification-policy Replaces the whole list of routes.

PUT /notification-policy body:

{
"routes": [
{
"name": "Critical to on-call",
"match": { "severity": ["critical"] },
"channels": ["<channel id>"],
"continue": true
},
{ "name": "Everything else", "match": {}, "channels": ["<channel id>"] }
]
}

Up to 100 routes, each with up to 20 channels. match takes severity, source, monitorIds, clientIds and labels. An empty match matches every alert.

Status Code When
400 UNKNOWN_CONNECTION The channel’s connection doesn’t exist.
400 NOT_A_NOTIFICATION_INTEGRATION The connection’s type can’t send notifications.
400 UNKNOWN_CHANNEL A route names a channel that doesn’t exist. details lists each one.

A connection holds the settings and secrets for one system SREInsights talks to. See Connections and secrets and Connection types.

Method Path What it does
GET /integrations The catalog: every connection type and what its form asks for. Needs a session, not an active organization.
GET /connections Connections.
POST /connections Creates a connection.
GET /connections/{id} One connection.
PATCH /connections/{id} Changes a connection.
POST /connections/{id}/ingest-token A new ingest token for a push source. The old one stops working. Shown once.
DELETE /connections/{id} Deletes a connection and its secrets.

POST /connections body:

Field Type Required Default What it does
type string Yes A type from GET /integrations.
name string Yes Up to 100 characters.
description string No Up to 500 characters.
clientId uuid No The client it belongs to. Leave out to share it across the organization.
runnerGroupId uuid No
secretMode sreinsights or runner No sreinsights runner: secrets never reach SREInsights; runners read them from SREINSIGHTS_SECRET_* variables.
config object No {} The type’s fields, secrets included.

Secrets never come back: the response’s secrets says only which are set. With secretMode: "runner", runnerVariables names the variable to set on each runner. A push source’s response has ingest.url, and its ingestToken (pit_…) once, on create.

In a PATCH, leaving a secret out keeps it; setting it to null removes it.

Status Code When
400 UNKNOWN_INTEGRATION No connection type with that name.
400 SECRET_MODE_UNSUPPORTED This type is used by SREInsights itself (notifications, push sources), so its secrets can’t stay on runners.
400 SECRET_REQUIRED A required secret was set to empty.
400 UNKNOWN_CLIENT, UNKNOWN_RUNNER_GROUP The client or runner group is not in this organization.
400 NOT_A_PUSH_SOURCE Only push sources have an ingest token.
409 CONNECTION_IN_USE Moving the connection to another client would break monitors or jobs that use it.

Push sources send their alerts to SREInsights. See Receive alerts from other tools.

Method Path What it does
POST /ingest/{type}/{connectionId} No user sign-in: authenticated by the connection’s ingest token. Answers 202 with { "received", "alerts" }.

type is kibana, dynatrace or alertmanager, and must match the connection’s type. Send the token as Authorization: Bearer pit_…. Copy the full URL from the connection.

Status Code When
400 INGEST_PAYLOAD_INVALID The body isn’t what this source sends.
401 INGEST_UNAUTHORIZED The token is missing or wrong, or no such connection of that type.

Data jobs collect data on a schedule with a built-in job type. See Jobs, Job types and Sync spec.

Method Path What it does
GET /job-types Every job type with its parameters, connection roles, default schedule and timeout.
GET /jobs Jobs.
POST /jobs Creates a job.
GET /jobs/{id} One job.
PATCH /jobs/{id} Changes a job. type can’t change.
DELETE /jobs/{id} Deletes a job.
POST /jobs/{id}/run Run now, with the job’s parameters overridden by params. Body { "params"? }. Answers 202.
POST /jobs/{id}/backfill One run for a range of days. Body { "start", "end", "params"? }, dates as YYYY-MM-DD, both included. Answers 202.
POST /jobs/{id}/preview Queues a dry run of a sync. Answers 202 with { "taskId", "state": "queued", "version", "queueExpiresAt" }.
GET /jobs/{id}/previews/{taskId} A preview’s state: queued, running, done (with what it found) or failed (with error).
GET /jobs/{id}/runs Runs, newest first. Takes limit and before.

POST /jobs body:

Field Type Required Default What it does
name string Yes Up to 120 characters.
description string No
clientId uuid or null No The client it belongs to. null for the organization’s own.
enabled boolean No true A sync must be created paused.
type string Yes A name from GET /job-types.
params object No {} The type’s parameters.
connections object No {} Up to 10 role: connection id pairs.
runsOn object Yes { "group", "labels": [] }, as for monitors.
schedule object Yes { "cron", "timezone": "UTC" }.
timeoutMs integer, 10000–900000 No The type’s default
alerting object No { "onFailure": true, "freshnessHours": null, "severity": "high" }.
sync object For type sync The sync’s spec.

PATCH /jobs/{id} also takes typeVersion, to move the job to another version of its type. It rejects fields it doesn’t know.

Run now and Backfill answer { "runId", "status": "queued", "queueExpiresAt", "params", "window" }.

Status Code When
400 UNKNOWN_JOB_TYPE No job type with that name, or the job’s version is no longer known.
400 BACKFILL_NOT_SUPPORTED The type collects from where it left off, so it has no day range.
400 NOT_A_SYNC Only a sync has a preview.
400 RUNNER_CREDENTIALS_NOT_HOSTED The type uses its runner’s own credentials, and the job runs on a hosted location.
400 HOSTED_RUNNER_LOCAL_SECRETS A connection keeps its secrets on your runners, and the job runs on a hosted location.
400 INVALID_SCHEDULE, SCHEDULE_TOO_FREQUENT, UNKNOWN_RUNNER_GROUP, UNKNOWN_CLIENT As for monitors.
400 VALIDATION_FAILED Also for a backfill range in the future or longer than the type allows in one run.
403 PLAN_LIMIT On hosted runners: the schedule is more frequent than your plan allows, or you are at your plan’s number of hosted jobs.
409 JOB_RUN_ACTIVE A run of this job is already waiting or running. A job runs one at a time.
409 SYNC_NOT_PREVIEWED The sync’s current version hasn’t been previewed. Save it paused, preview it, then switch it on or run it.

Clients group monitors, jobs, connections and status pages inside one organization. See Clients.

Method Path What it does
GET /clients Clients.
POST /clients Creates a client. Body { "name" }.
PATCH /clients/{id} Renames a client. Body { "name" }.
DELETE /clients/{id} Deletes a client that nothing belongs to any more.
Status Code When
409 CLIENT_NAME_TAKEN A client with this name exists.
409 CLIENT_IN_USE Records or members still belong to the client. Move or delete them first.

See Status pages and Share a status page.

Method Path What it does
GET /status-pages Status pages. ?clientId=<id> for one client’s, ?clientId=none for the organization’s own.
POST /status-pages Creates a page.
GET /status-pages/{id} One page.
GET /status-pages/{id}/summary Uptime, per-monitor rates and certificates. ?window=day, week (default) or month.
PATCH /status-pages/{id} Changes a page.
PUT /status-pages/{id}/visibility Owners and Admins only. Body { "visibility": "private" | "link" }.
DELETE /status-pages/{id} Deletes a page.
GET /public/status/{token} Public, no sign-in. A page shared by link.

POST /status-pages body: name, description, clientId (null for the organization’s own), and items, up to 50:

{
"name": "Acme services",
"items": [
{ "monitorId": "<monitor id>", "label": "Checkout" },
{ "monitorId": "<monitor id>", "kind": "certificate" },
{ "kind": "job", "jobId": "<job id>", "label": "Nightly import" }
]
}

A monitor item’s kind is monitor (default) or certificate (only its TLS certificate).

Setting visibility to link makes a new share token each time. Owners and Admins see it as shareToken; the page’s address is <your SREInsights URL>/s/<token>. Setting it back to private turns the old address off.

GET /public/status/{token} answers a plain-text 404 for a wrong or turned-off token. Answers may be cached for 60 seconds, carry x-robots-tag: noindex, nofollow, and are rate limited (see Rate limits).

Status Code When
400 UNKNOWN_MONITOR, UNKNOWN_JOB An item names a monitor or job that doesn’t exist.
400 ITEM_OF_ANOTHER_CLIENT A client’s page lists another client’s monitor or job.
400 UNKNOWN_CLIENT clientId is not in this organization.
403 POLICY_DENIED Changing visibility without the Owner or Admin role.

Scheduled summaries of monitors, sent to notification channels. See Reports.

Method Path What it does
GET /reports Reports.
POST /reports Creates a report.
GET /reports/{id} One report.
GET /reports/{id}/runs The report’s past sends.
PATCH /reports/{id} Changes a report.
POST /reports/{id}/send Sends it now. Answers 202.
DELETE /reports/{id} Deletes a report.

POST /reports body:

Field Type Required Default What it does
name string Yes Up to 120 characters.
description string No
enabled boolean No true
cron string Yes When to send.
timezone string No UTC
monitorIds uuid[] No [] Up to 100.
channelIds uuid[] Yes 1–20 notification channels.
format summary or detailed No summary
includeRuns boolean No true Count every run since the last report, not only the latest result.

Sending a report with no channels answers 400 REPORT_HAS_NO_CHANNELS.

A deployment is a planned change with its checklist, approvals and tickets. See Deployments.

Method Path What it does
GET /deployments Deployments. Filter with ?status= and ?environment=.
POST /deployments Creates a deployment.
GET /deployments/{id} One deployment.
PATCH /deployments/{id} Changes a deployment. checklist, approvals and tickets are replaced as a whole.
POST /deployments/{id}/duplicate Copies a deployment. Answers 201.
DELETE /deployments/{id} Deletes a deployment.
Field Type Default What it does
name string, required Up to 160 characters.
description string Up to 4,000 characters.
environment development, staging, production production
status draft, pending_approval, approved, in_progress, completed, failed draft
scheduledAt ISO date-time or null
checklist array, up to 100 [] Items with title, description, done, doneAt, doneBy, ticket, approvals.
approvals array, up to 50 [] { "name", "email", "status": "pending" | "approved" | "rejected", "comment", "respondedAt" }.
tickets array, up to 50 [] { "key", "title", "url" }, for example JIRA-123.

ADAM, the AI assistant. See ADAM. The operator must give the instance an AI API key; until then, /ai/flows, /ai/analyze, /ai/scripts, /ai/tests and /ai/fix answer 503 AI_DISABLED.

Method Path What it does
GET /ai/status { "enabled", "model" }.
POST /ai/flows A flow from a description. Body { "prompt", "current"? }.
POST /ai/analyze Reads a ticket or incident text: entities, errors and suggested checks. Body { "text" }.
POST /ai/scripts Code for a script step. Body { "prompt", "language": "node" | "python" | "shell", "current"?, "runtime"?, "inFlow"?, "stepIds"?, "envVars"? }.
POST /ai/tests A monitor as test code. Body { "name", "description"?, "definition", "parameters"?, "framework": "playwright" | "pytest" | "k6" | "postman" }.
POST /ai/fix A patch for pasted code. Body { "incident", "code", "filename"?, "language"? }.
POST /ai/code/fetch Asks one of your runners for excerpts of the files an incident names. Body { "connectionId", "groupId", "repository"?, "ref"?, "text"? or "hints"? }. Answers 202 with { "taskId", "hints" }.
GET /ai/code/tasks/{id} A code task’s state: queued, running, done or failed.
Status Code When
422 AI_REFUSED The AI declined the request.
422 NO_CODE_HINTS No file names found in the text. Add hints by hand.
422 NOT_A_CODE_HOST The connection is not GitHub or GitLab.
422 HOSTED_GROUP_NOT_ALLOWED Find the code runs only on your own runner groups.
422 BAD_REPOSITORY Name the repository as owner/repo (GitHub) or group/project (GitLab).
429 AI_RATE_LIMITED The AI service is limiting requests. Try again in a minute.
502 AI_AUTH_FAILED, AI_UNREACHABLE, AI_FAILED, AI_UNPARSEABLE The AI service rejected the key, couldn’t be reached, failed, or gave an answer that couldn’t be read.
503 AI_DISABLED ADAM is off on this instance.

Connect Etsy and eBay stores by signing in to them. Available only where the operator set it up. See Stores.

Method Path What it does
GET /stores The providers (with whether each is on) and the connected stores.
POST /stores/connect Starts a sign-in. Body { "provider": "etsy" | "ebay", "storeKey", "clientId"?, "label"? }. Answers 201 with { "authorizeUrl", "expiresAt" }.
POST /stores/callback Finishes a sign-in with what the provider sent back. Body { "provider", "state", "code"?, "error"?, "errorDescription"? }.
DELETE /stores/{id} Forgets the store in SREInsights. Its stored token and the provider’s approval stay.

Open authorizeUrl in the browser of the person who started the sign-in. It works for ten minutes. storeKey is lower-case letters, digits and single hyphens.

Status Code When
400 STATE_EXPIRED The sign-in took more than ten minutes. Start again.
400 STATE_UNKNOWN The sign-in is unknown or was for another provider. Start again.
404 STORE_CONNECT_OFF Connecting stores isn’t set up for this organization on this instance.
409 STORE_OWNED_ELSEWHERE The store is connected in another organization on this instance.
409 STORE_LOCKED The store’s token is being refreshed right now. Wait a minute and try again.
409 STORE_UNKNOWN The store isn’t known to the instance, or was removed while you signed in.
Method Path What it does
GET /insights/overview Runs per day and this month’s totals, for the Overview page. ?days= 1–31, default 7.

Two endpoints sit outside /api/v1 and need no sign-in:

Method Path What it does
GET /healthz { "status": "ok" } while the server process runs.
GET /readyz { "status": "ready" } when the database is reachable, otherwise 503.

Set these once:

Terminal window
export SRI=https://console.sreinsights.com/api/v1
Terminal window
curl -s -c cookies.txt -X POST "$SRI/auth/login" \
-H 'content-type: application/json' \
-d '{"email": "you@example.com", "password": "your password"}'
{ "status": "success", "data": { "accessToken": "eyJhbGciOiJIUzI1NiIs…" } }

Keep the access token, and the cookie file for refreshing:

Terminal window
export TOKEN=eyJhbGciOiJIUzI1NiIs…
curl -s -b cookies.txt -c cookies.txt -X POST "$SRI/auth/refresh"
Terminal window
curl -s "$SRI/monitors" -H "authorization: Bearer $TOKEN"
{
"status": "success",
"data": [
{
"id": "0193a1c2-…",
"name": "Checkout API",
"enabled": true,
"runsOn": { "groupId": "0193…", "group": "eu-prod", "hosted": false, "labels": [] },
"schedule": { "cron": "*/5 * * * *", "timezone": "UTC", "inputs": {}, "nextRunAt": "2026-11-22T10:05:00.000Z", "lastRunAt": "2026-11-22T10:00:00.000Z" },
"lastStatus": "ok",
"activeRun": null
}
]
}
Terminal window
curl -s -X POST "$SRI/monitors/0193a1c2-…/run" \
-H "authorization: Bearer $TOKEN" \
-H 'content-type: application/json' \
-d '{"params": {"region": "eu"}}'
{
"status": "success",
"data": { "jobId": "0193b7d0-…", "status": "queued", "queueExpiresAt": "2026-11-22T10:08:12.000Z", "params": { "region": "eu" } }
}

Watch for its result with GET /monitors/{id}/results?limit=1.

Use the URL and token you got when you created the trigger:

Terminal window
curl -s -X POST "https://console.sreinsights.com/api/v1/hooks/0193c4e1-…" \
-H "authorization: Bearer pwt_…" \
-H 'content-type: application/json' \
-d '{"params": {"version": "2.4.1"}, "commit": "a1b2c3d"}'
{ "status": "success", "data": { "queued": true, "jobId": "0193c4f0-…", "params": { "version": "2.4.1" } } }

The whole body, commit included, is available to the check as trigger.body.