TRCR dashboards

Grafana dashboards for TRCR, network monitoring you run yourself. TRCR agents probe your targets from several places at once (ping, HTTP, DNS, SSL, BGP, traceroute, throughput and agent-to-agent mesh) and write the results to your own PostgreSQL or InfluxDB. These dashboards read that data.

Download them, point them at your datasource, and change them to suit your own TRCR deployment.

One bad day, seen from four places

Four agents, in Dallas, Los Angeles (two) and San Francisco, watch the same targets. For three days nothing happens. Then, at 08:45 one morning, latency rises on every long path at once, stays high for about ten hours, and returns to normal at 19:30. Here is how that day looks in TRCR.

1. Everything in one place

The Ping dashboard shows round-trip time and packet loss for every agent and target, and you can filter by either one. Over three days, a single block stands out.

2. Three quiet days, then a step

Paths that sat at 32 ms jump to between 46 and 68 ms. It isn't a spike or a slow drift: the change is sudden and it holds.

3. The shape of the incident

Zoomed to 36 hours: the rise at 08:45, a slow climb through the afternoon on the worst paths, and a clean recovery at 19:30.

4. Loss follows latency

Packet loss over the same 36 hours. Bursts of up to 25% appear on the slowest paths and cluster between 15:00 and 19:30, as the latency climbs.

5. Is it the target, or the path?

Cross-vantage spread is the gap between the fastest and the slowest agent reaching the same target. When it doubles, from about 32 ms to about 65 ms, the agents disagree, so the trouble is on some paths and not at the target.

6. What was not affected

Minimum, average and maximum round-trip time for every agent and target pair. The long paths at the top show the incident. The short ones, at 13 ms, 9 ms and under 1 ms, never move, which narrows down where the problem was.

7. What users felt

HTTP response time from the user's side, as a daily median (p50) and 95th percentile (p95) over three and a half weeks. This is the view to check when you need to know whether a network event reached your customers.

8. The agents watch each other

Every agent also probes every other agent. With four agents in a full mesh there are 12 links, and all 12 are up.

9. A baseline between your own sites

The latest probe on each mesh link: half a millisecond between the two Los Angeles agents, 9 ms from Los Angeles to San Francisco, 32 ms from Dallas to Los Angeles and 41 ms from Dallas to San Francisco. When a number here changes, the network between your own sites changed.

The dashboards

There are two editions with the same layout. Use the one that matches where your TRCR master writes its results.

InfluxDB v2 (Flux)

Datasource uid trcr-influxdb. Files in dashboards/influxdb/.

Dashboard	What it shows
Agent Mesh	Agent-to-agent mesh — same shape as the Postgres edition, queries InfluxDB v2 via Flux.
Per-Agent Operations	Fleet health from the agent perspective — probes/min, success rate, last-seen, top failing targets.
BGP	BGP route lookup + GLOBAL AS-path snapshot from RIPEstat's RIS route collectors (~23 vantage points worldwide).
Cert Fleet	Glanceable certificate-fleet view — one threshold-coloured tile per ssl_expiry probe, soonest expiry first, plus the fleet timeline.
DNS Propagation	Multi-resolver DNS propagation + divergence.
DNS	Single-resolver DNS query latency, answers, rcode.
Executive Scorecard	Is the network healthy and trending the right way, in 30 seconds.
HTTP	HTTP probe latency + status.
Infrastructure Hygiene	The quiet correctness layer: time sync, path MTU, and port reachability.
Map	Geographic up/down view of every probe target.
NOC Wallboard	Everything currently wrong, one screen.
Path & Latency Forensics	Overlay what changed on when it changed: RTT, loss, spread, path and AS-path events for one target.
Ping	Ping (ICMP) probe latency + loss, per agent → target.
Risk & Lifecycle Outlook	Slow-moving risks before they page anyone: certs, crypto, routing churn, capacity, clock drift.
Service Journey	One service, whole delivery stack, every vantage point.
SLA & Availability Report	Defensible per-service availability against SLO.
SLO Landing	Default 'is everything OK?' view.
SSL Chain Health	Full-chain SSL health — earliest expiry across leaf + every intermediate, CA concentration, chain depth, per-target chain detail.
SSL Expiry	SSL/TLS certificate expiry.
Throughput	TRCR-native TCP throughput probe (replaces iperf3 shell-out).

PostgreSQL

Datasource uid trcr-postgres. Files in dashboards/postgres/.

Dashboard	What it shows
Agent Mesh	PerfSonar-style agent-to-agent mesh.
Per-Agent Operations	Fleet health from the agent perspective — probes/min, success rate, last-seen, top failing targets.
Agent-to-Agent	Agent-to-agent connectivity matrix.
Alive	Alive (ARP/ND/TCP-knock) sweep results — per-device liveness across a fleet.
BGP	BGP route lookup + GLOBAL AS-path snapshot from RIPEstat's RIS route collectors (~23 vantage points worldwide).
Cert Fleet	Glanceable certificate-fleet view — one threshold-coloured tile per ssl_expiry probe, soonest expiry first, plus the fleet timeline.
DNS Propagation	Multi-resolver DNS propagation + divergence.
DNS	Single-resolver DNS query latency, answers, rcode.
HTTP	HTTP probe latency + status.
Map	Geographic up/down view of every probe target.
Overview	TRCR v2 fleet-wide overview: probe volume, success rate, recent failures, distribution by probe_type.
Ping	ICMP ping RTT, loss, and jitter per target.
Cross-Vantage RTT Divergence	For each target, the spread (max − min) of average ICMP RTT across all enrolled agents.
SLO Landing	Default 'is everything OK?' view.
SSL Expiry	SSL/TLS certificate expiry.
Throughput	TRCR-native TCP throughput probe (replaces iperf3 shell-out).
Traceroute	Traceroute hop counts and path-change detection.

Installing

1. Add the datasource. The dashboards find their datasource by uid, so the uid has to match: trcr-influxdb or trcr-postgres.
   • InfluxDB: copy datasources.yaml.example to /etc/grafana/provisioning/datasources/, put in your URL, organisation, bucket and a read-only token, and restart Grafana. The datasource must use Flux, not InfluxQL.
   • PostgreSQL: add a PostgreSQL datasource with uid trcr-postgres that points at your TRCR results database. A read-only role is enough.
2. Add the dashboards. Import the JSON files in Grafana under Dashboards → New → Import, or copy them to your dashboards provisioning folder and restart Grafana.

Never put a real token or password in a file you share. Keep them in your own provisioning files on the Grafana host.

Contributing

This repository is public. Before every commit and push, a check looks for keys, passwords, tokens, public IP addresses, email addresses and image metadata, and reads the text in images to catch IP addresses in screenshots. Turn it on once after cloning:

git config core.hooksPath .githooks

Screenshots must not show real IP addresses, hostnames or account names. Crop or re-capture them first. scripts/check-public.sh --all checks everything.

Licence

Copyright © 2026 TRCR Networks. The dashboards are free to use and modify with your own TRCR deployment. They may not be resold, republished elsewhere or used with a competing product. The full terms are in LICENSE.