Connected monitoring

One service, one place: endpoint checks, its metrics and its dashboards together.

What it is#

One service, one place. An endpoint check answers "can users reach it right now?" (outside-in). Its metrics answer "what is happening inside?" (inside-out). Pulsimo links the two, and the dashboards that describe them, so "it's slow" becomes "it's slow because p95 on /checkout jumped and the database pool is full", on one screen.

Where to find it#

  • Services → Endpoints → click a card → Metrics & dashboards tab.
  • Endpoint cards (Compact or Detailed, switch above the grid).
  • Metrics → Sources → Connect for each source.
  • Dashboards: "Monitors:" chips under the title.

Linking an endpoint to its metrics#

Pulsimo suggests links; you confirm with one click.

Suggestion Example
Same route endpoint /api/v1/orders ↔ route="/api/v1/orders" (also path, handler, uri…)
Same host endpoint http://10.0.1.5:8080/... ↔ a source on 10.0.1.5:8080
Same database a PostgreSQL check ↔ the postgres_exporter for that server
Same project the source and the endpoint are in the same project

You can also link a metric source manually with your own label filters. One endpoint can link to several sources: the app, its database, the host it runs on.

What a link gives you depends on what the source exposes (its profile):

Profile Charts and numbers
HTTP app requests/s, 5xx share, p95 latency (for the linked route)
Process CPU, memory
Host (node_exporter) CPU, memory
PostgreSQL / MySQL up, connections
Containers (cAdvisor) CPU, memory

The Metrics & dashboards tab#

From top to bottom:

  1. How it's monitored: Active check, From metrics, or Both (Monitoring from metrics).
  2. Reliability target: set or view the endpoint's SLO (SLOs).
  3. Suggested links, one click each.
  4. Charts per linked source for 1h, 6h, 24h or 7d, with Explore on each chart. Dashed vertical lines mark deploys and restarts of those sources (Change markers).
  5. Dashboards: attach an existing one (variables such as route and source are preset when you open it from here) or Create dashboard for .

Endpoint cards#

Linked cards show live numbers from the metrics (4.2 req/s · 0.3% 5xx · p95 38 ms). A value over its threshold turns amber or red, even while the check itself is Up, which is exactly the early warning you want.

From a source to monitors#

Metrics → Sources → Connect:

  • Monitor API routes (app sources): the routes found in the request metrics, with req/s and methods. Templated routes such as /users/{id} ask for a real value for an active check. Creates the checks already linked.
  • Monitor this database (PostgreSQL/MySQL sources): opens the add-endpoint form pre-filled and links the new check to the exporter.
  • Create dashboard.

The Discovered inbox (Discovered services) does the same in the background, for every source.

In alerts and incidents#

  • "Service down" and escalation alerts include What the metrics show, read at send time ("p95 latency 2.1 s (was 40 ms 30 min ago)", "orders-db: 100% connections used").
  • Incidents show the linked metrics around the start, saved with the incident (Incidents and evidence).