v2.3.0

This release adds a service_info gauge to the Prometheus metrics, roughly analogous to other kinds of e.g. build_info gauges, for every configured or discovered service.

service_info{service_id="XXX", service_name="YYY", service_version="123"} 1

v2.2.1

Minor updates, including one potential fix for a resource leak.

v2.2.0

Minor improvement release: if the Fastly API -token is invalid (for example, by expiring) the monitoring goroutine will no longer slam rt.fastly.com in a hot-loop.

v2.1.0

• Minimal (~11MB) Dockerfile container build.
• Available as mrnetops/fastly-exporter from Docker Hub.
• Set default namespace & subsystem (fastly and rt respectively).
• Read token from environment variable (FASTLY_API_TOKEN).
• Docker Compose for a full fastly-exporter + Prometheus +
  Grafana + Fastly dashboard stack.
• Fix a few typos in metric help messages.

v2.0.0

• Service IDs are now optional. By default, the fastly-exporter will poll the Fastly API for every service accessible to your API Key (-token) and provide metrics for all of them.
• The individual tls_v1x_total metrics have been collapsed into the tls_version label on the tls_total metric. The existing tls_total metric has been given the label "all".
• It seems like people are actually using this thing (weird) so I've added tests and continuous integration.

v1.0.1

Adds counters for TLS 1.0, 1.1, and 1.3 hits, to accompany TLS 1.2, which already existed. See PR #7.

v1.0.0

The exporter now adds a service_id label with the (immutable) service ID to each metric. Breaking change: for consistency and explicitness, the (mutable) service name is now reported under the label service_name, where previously it was service. See PR #6.

v0.0.3

Add support for multiple services (issue #2, PR #4) by repeating the -service flag.

v0.0.2

Fix a dumb crashing bug.

v0.0.1

Initial release.