One agent per host. Auto-instrumented, queryable in plain SQL, and priced per host rather than per gigabyte you happen to send us.
No card required ·
Example customer profiles
The agent discovers services on the host, instruments the runtimes it recognises, and starts baselining. Nothing to add to your application code.
A single binary, no sidecar per service, no forked SDK to keep in step with upstream. Auto-instrumentation covers Go, Node, Python, Java, Ruby and .NET.
# install and point it at your cluster curl -sSL pulsegrid.example/install | sh pulse connect --env production -> 41 services discovered, tracing live
Keeps every errored and slow trace, thins the successful ones.
85% less stored, error traces kept in full
Every deploy, feature flag change, config push and scaling event is drawn on the same timeline as your latency graph, so a step change lines up with whatever caused it.
Traces and logs are stored in columnar tables you can select from. No proprietary query language, and you can point any BI tool at the same read replica.
Thresholds learned per service, per hour of the week. Sunday traffic does not page you.
Self-host the full stack under a source-available licence, or run managed in any of five regions. Export to Parquet at any time. There is no egress charge for taking your data out, including on the way to a competitor.
Per host, per month. Ingest, retention and queries are included at every tier. There is no per-gigabyte charge and no charge for running a query, which are the two lines that make observability bills unpredictable.
Nothing is dropped. You are billed for the extra hosts on the next invoice at the same per-host rate, pro rata from the day they appeared. We email the account owner when host count rises more than 20% above the previous month.
We accept OpenTelemetry over OTLP, so if you are already exporting OTel you can point it at us and run both in parallel. Dashboards do not migrate automatically. Most teams rebuild the ten dashboards they actually use and discard the rest.
99.9% on Team, 99.95% on Scale, measured monthly on the ingest and query APIs. Credits are 10% of monthly fee per 0.1% below target, applied automatically without you having to claim.
Four minutes to install, no card, and the free tier covers three hosts indefinitely. Break something on purpose and see whether the timeline tells you what you already know.