Service health
Insights → Service health shows whether Merge Pilot can reach the systems it works with, and whether its notifications get through. GitLab and Jira connection checks for every company. Refreshes every 30 seconds.
People of a company see their own company only; platform admins see every company.
Connection checks
Merge Pilot checks every active VCS connection and every Jira connection of the active companies about once a minute, the same way as Test connection (see Test connection). The built-in tracker has nothing to reach and is not checked.
The cards at the top count Total services, Healthy and Unhealthy. The table has one row per connection:
| Column | Meaning |
|---|---|
| Service | The provider (GitLab, GitHub, Jira) and the connection's name: a company with two connections of one provider has a row for each |
| Company | The company the connection belongs to |
| Status | Healthy or Unhealthy, from the latest check |
| Latency | How long the latest check took |
| Uptime (24h) | The share of the last 24 hours' checks that passed |
| Last check | When the latest check ran |
| Error | Why the latest check failed |
Before the first check of a connection the page says No health checks yet. When a refresh of the page fails, it keeps the last results and says The latest refresh failed (…); the results below may be out of date.
A check that finds the VCS refusing Merge Pilot's credentials, a GitHub App missing permissions or the rate limit used up also records the problem on the connection and announces it (see Connection health). For a GitHub App connection the check also calls GitHub with each installation's token; a problem is cleared only by a check that used the credentials that failed (see How connection health is decided).
Running instances
Platform admins also see a card N running instances: the servers that make up this Merge Pilot installation, each with its name, version, when it started and when it last reported in. Each one reports every few seconds; one that has not reported for three of those is no longer listed, and one that is stopped leaves the list at once. The request limits (see Rate limits) are for all of them together. Company people do not see the card.
Notification deliveries
When notification deliveries failed for good in the last 7 days, a Notification deliveries card lists each company concerned: <company>: N notifications could not be delivered in the last 7 days. Company admins and platform admins get Open the delivery log, which opens the company's Notifications tab. See Delivery log and failures.
API
| Request | Does |
|---|---|
GET /api/v1/health-checks | The recorded checks, filtered by service, is_healthy, from, to, paged |
GET /api/v1/health-checks/summary | The latest state of each connection (connection_id, connection_name) |
GET /api/v1/health-checks/notifications | The companies with failed notification deliveries in the last 7 days: company_id, company_slug, company_name, failed |
Platform admins choose a company with company_id; others always get their own company.
Related pages
- Connections: test connection and connection health.
- Delivery log and failures: failed notifications.
- Jobs: background work and its failures.
Video tutorial
Short videos of this feature working on real GitLab and Jira projects. English audio, English and Turkish subtitles. All video tutorials →