Sapient Pulse (telemetry)
Sapient Pulse is the monitoring built into Sapient’s own hosting. Products deployed with Sapient Deploy export to it, and what they export flows back into your factory — no Datadog account, no agent to install, no dashboards to build.
What it enables: your seats can answer “is it up, and what broke” about a product they shipped, using real data from the running app.
What it collects
Section titled “What it collects”| Signal | What you get |
|---|---|
| Logs | What your product emits through OpenTelemetry, searchable by service and time. |
| Traces | Request timings — which call was slow, and where it went. |
| Metrics | Request rate, error rate, and latency percentiles per service. |
| Errors | Failing requests, given a far larger quota than ordinary records so a failing product stays visible under load. |
| Health | A per-service verdict: healthy, degraded, unhealthy — or no data. |
That last one matters more than it looks. A product with no traffic reports “no data”, never “healthy”. A green light nothing has verified is worse than no light, so Pulse doesn’t show one.
Where it shows up
Section titled “Where it shows up”- Observability in the control room — logs, traces and metrics for any hosted product.
- After every deploy — one health check reads Pulse and, if the new release is already failing, raises a rollback recommendation for you to approve. A product with no traffic yet reports no data, which never triggers one.
- Your SRE and engineering seats — they query it while investigating, the same way they’d query a connected Datadog or Cloud Logging.
What your product has to do
Section titled “What your product has to do”Every deploy is handed the standard OpenTelemetry environment variables, so a product’s exporter is already pointed at Pulse and authenticated. The product itself still has to be instrumented — that’s a build obligation your engineering seats are given as part of their guidance, and they do it as a matter of course. There is no configuration step for you, but the data you see is whatever the app actually exports.
Scope and limits
Section titled “Scope and limits”Pulse covers products hosted on Sapient. It is retention-limited (roughly a week by default) and rate-capped per factory, per signal, per minute — so a very chatty product can shed records, with errors given a far larger allowance than the rest. Most queries that hit their scan ceiling say so, with a “results may be partial” note rather than a silently short answer. It has no alerting, no custom dashboards, and no on-call routing.
Products that need those should connect a real observability provider through the ordinary connector flow; the factory reads whichever one is connected through the same screens.
Reading it costs credits
Section titled “Reading it costs credits”Log and APM queries are metered like any other read — see Billing and credits. Ingest is not billed.