# ctlog-uptime-exporter

![grafana dashboard](grafana.png)

A Prometheus exporter for [Certificate Transparency](https://certificate.transparency.dev/) log uptime data published by Google at:

```
https://www.gstatic.com/ct/compliance/endpoint_uptime_24h.csv
```

The CSV reports the 24-hour uptime percentage for each CT log URL broken down by operation type (`add-chain`, `get-entries`, `get-sth`, etc.). This exporter fetches that CSV on a configurable schedule and exposes the data as Prometheus metrics.

## Metrics

| Metric | Labels | Description |
|---|---|---|
| `ct_log_uptime_ratio` | `log_url`, `endpoint` | 24h uptime as a ratio (0-1) |
| `ct_log_uptime_fetch_success` | | 1 if the last fetch succeeded, 0 otherwise |
| `ct_log_uptime_fetch_timestamp_seconds` | | Unix timestamp of the last fetch attempt |

Standard `go_*` and `process_*` metrics are also exposed.

## Usage

```sh
go build -o ctlog-uptime-exporter .
./ctlog-uptime-exporter
```

Metrics are served at `http://localhost:9781/metrics`.

## Flags

| Flag | Default | Description |
|---|---|---|
| `-listen` | `:9781` | Address to listen on |
| `-url` | `https://www.gstatic.com/ct/compliance/endpoint_uptime_24h.csv` | URL of the uptime CSV |
| `-interval` | `12h` | How often to fetch the CSV |
| `-jitter` | `5m` | Maximum +/-jitter applied to the fetch interval |

The exporter fetches the CSV once on startup, then repeats every `-interval` +/- a random value up to `-jitter`. This avoids thundering-herd effects if multiple instances run in parallel.

## Example Prometheus scrape config

```yaml
scrape_configs:
  - job_name: ctlog_uptime
    static_configs:
      - targets: ['localhost:9781']
```

## systemd

A unit file and defaults file are included.

```sh
# install binary
go build -o /usr/local/bin/ctlog-uptime-exporter .

# install defaults (edit to taste)
cp ctlog-uptime-exporter.default /etc/default/ctlog-uptime-exporter

# install and start the service
cp ctlog-uptime-exporter.service /etc/systemd/system/
systemctl daemon-reload
systemctl enable --now ctlog-uptime-exporter
```

Runtime flags are controlled via the `ARGS` variable in `/etc/default/ctlog-uptime-exporter`.

## Grafana dashboard

`dashboard.json` can be imported directly into Grafana (Dashboards -> Import).
It expects a Prometheus datasource and provides:

- Summary stats: number of logs, endpoint types, average uptime, degraded count, fetch status, last fetch time
- Variable selectors for Log URL and Endpoint (both multi-select with All)
- Time series panel showing the rolling 24h uptime ratio over the chosen time range
- Table of the top N least-available log/endpoint pairs (N is selectable: 5, 10, 25, 50)

## License

Apache 2.0 - see [LICENSE](LICENSE).