SrvPulse
regional collector

Monitoring that survives the network between you and it.

Agents report to a collector inside their own region. The collector batches, compresses and forwards upstream over a single long-lived channel — so a flaky or restricted path between regions costs you latency, never data.

This host

Endpoint
api.srvpulse.net
Role
regional collector
Region
ru-central
Location
Moscow, RU
Upstream
connect.api.srvpulse.net
Port
443 (TLS 1.2+)
API
v1

Regional collectors terminate agent connections inside the local network and forward aggregated batches upstream over a single persistent channel. See collector topology.

Agentsyour hosts
→
Regional collectorapi.srvpulse.net
→
Aggregationconnect.api.srvpulse.net
9collector regions
10sminimum interval
99.97%ingest availability, 90d
13 moretention on Scale

Why a collector tier

Sending every sample from every host straight to a central API works until the path between your region and ours degrades. Then you lose exactly the data you need to explain the incident. A regional collector keeps the hot path short and local:

  • Agents never cross a border. They talk to a host in their own region, over a connection that stays inside the local network.
  • One channel upstream instead of thousands. The collector multiplexes every agent in the region into a single authenticated connection to aggregation.
  • Buffered, not dropped. If the upstream channel is unavailable, the collector spools to disk and replays oldest-first. Batches are idempotent, so a replay never duplicates.
  • Local queries stay local. Recent windows are answerable from the collector without a round trip to aggregation.

What it watches

CheckCollectedStatus
Host metricsCPU, memory, disk, IO, loadGA
Process & systemd unitsstate, restarts, RSSGA
HTTP / TCP probeslatency, status, TLS expiryGA
Container runtimeDocker, containerdGA
PostgreSQL / Redisconnections, replication lagGA
Log tailingmatched patterns onlybeta