Monitoring Tools

Tools for monitoring and analyzing Fediverse instances.

Instance Directories

FediDB

Fediverse statistics and analytics.

Property	Value
Website	fedidb.org
Features	Stats, software tracking
Data	User counts, posts, growth

instances.social

Mastodon instance directory.

Property	Value
Website	instances.social
Features	Discovery, filtering
Focus	Mastodon instances

Server Monitoring

UptimeRobot

Basic uptime monitoring.

Property	Value
Website	uptimerobot.com
Free tier	50 monitors (pricing page, October 2026)
Checks	HTTP, keyword

Setup for Fediverse:

Monitor: HTTP(s)
URL: https://your-instance.example/api/v1/instance
Keyword: "uri"

Prometheus + Grafana

Advanced metrics collection.

Mastodon metrics: Mastodon has a built-in Prometheus exporter, enabled with MASTODON_PROMETHEUS_EXPORTER_ENABLED=true (default port 9394). See the Mastodon configuration docs for the full set of MASTODON_PROMETHEUS_EXPORTER_* variables.

# prometheus.yml
scrape_configs:
  - job_name: 'mastodon'
    static_configs:
      - targets: ['localhost:9394']

Health Checks

Endpoint monitoring:

# Check instance health and public stats
curl -s https://instance.example/api/v1/instance | jq '.stats'

Federation Monitoring

Federation Status

Check federation health:

# Test outgoing federation
curl -X POST https://your-instance.example/api/v1/statuses \
  -H "Authorization: Bearer $TOKEN" \
  -d "status=Federation test $(date)"

For Mastodon's delivery queues, use the Sidekiq web UI at /sidekiq (admin only) or the Sidekiq API below.

Blocked Instances

Track moderation actions:

# Get domain blocks — public only if the admin chose to show them
# (Mastodon's show_domain_blocks setting); otherwise returns 404 or needs a login
curl https://instance.example/api/v1/instance/domain_blocks

Queue Monitoring

Sidekiq (Ruby platforms)

# Check queue status
require 'sidekiq/api'

stats = Sidekiq::Stats.new
puts "Processed: #{stats.processed}"
puts "Failed: #{stats.failed}"
puts "Queued: #{stats.enqueued}"

Log Analysis

Common Patterns

Mastodon logs to stdout. On a standard install the systemd units send this to the journal; with Docker use docker compose logs.

Signature failures:

journalctl -u mastodon-web --since "1 hour ago" | grep -i "signature"

Federation errors:

journalctl -u mastodon-sidekiq --since "1 hour ago" | grep -iE "delivery|error"

Log Aggregation

Using Loki + Grafana:

# promtail config
scrape_configs:
  - job_name: mastodon
    static_configs:
      - targets:
          - localhost
        labels:
          job: mastodon
          # Mastodon logs to the systemd journal; for journald use
          # promtail's `journal` scrape config instead of __path__

Performance Tools

Database Monitoring

PostgreSQL queries:

-- Active connections
SELECT count(*) FROM pg_stat_activity;

-- Long-running queries
SELECT pid, now() - pg_stat_activity.query_start AS duration, query
FROM pg_stat_activity
WHERE state = 'active' AND now() - pg_stat_activity.query_start > interval '5 seconds';

-- Table sizes
SELECT relname, pg_size_pretty(pg_total_relation_size(relid))
FROM pg_catalog.pg_statio_user_tables
ORDER BY pg_total_relation_size(relid) DESC;

Redis Monitoring

# Memory usage
redis-cli INFO memory

# Key count
redis-cli DBSIZE

# Slow log
redis-cli SLOWLOG GET 10

Alerting

PagerDuty / Opsgenie

Set up alerts for:

• Instance down
• High error rate
• Queue backup
• Database issues

Example Alert Rules

# Prometheus alert rules
groups:
  - name: fediverse
    rules:
      - alert: InstanceDown
        expr: up{job="mastodon"} == 0
        for: 5m
        annotations:
          summary: "Instance is down"

      - alert: QueueBacklog
        expr: max by (queue) (ruby_sidekiq_queue_backlog) > 10000   # needs SIDEKIQ_DETAILED_METRICS
        for: 10m
        annotations:
          summary: "Sidekiq queue is backing up"

Dashboard Example

Grafana dashboard panels:

Panel	Query	Purpose
Response Time	ruby_http_request_duration_seconds	Performance (needs MASTODON_PROMETHEUS_EXPORTER_WEB_DETAILED_METRICS=true)
Error Rate	rate(ruby_http_requests_total{status=~"5.."}[5m])	Reliability (same)
Queue Depth	max by (queue) (ruby_sidekiq_queue_backlog)	Processing (needs MASTODON_PROMETHEUS_EXPORTER_SIDEKIQ_DETAILED_METRICS=true)
Queue Latency	max by (queue) (ruby_sidekiq_queue_latency_seconds)	Delivery delay (same)
Failed Jobs	rate(ruby_sidekiq_failed_jobs_total[5m])	Delivery failures (same)

Per-queue metrics are reported by every Sidekiq process that handles the queue, so take the max, don't sum (Mastodon's sidekiq.rb initializer). Without the detailed-metrics flags, Mastodon still exports Ruby process, Sidekiq process (busy/concurrency) and ActiveRecord connection-pool metrics.

Mastodon exports no product metrics (users, posts) — these come from the prometheus_exporter gem; check your own /metrics output for exact names.

See Also

• Testing Tools
• Server Software Overview
• Hosting Providers