Managing Apache Druid (Deep)

Manage and monitor Apache Druid cluster — datasources, ingestion, segments, queries, and cluster health.

Discovery Phase

#!/bin/bash

DRUID_API="${DRUID_ROUTER_URL:-http://localhost:8888}"

echo "=== Cluster Health ==="
curl -s "$DRUID_API/status" \
  | jq '{version: .version, memory: .memory}'

echo ""
echo "=== Datasources ==="
curl -s "$DRUID_API/druid/coordinator/v1/datasources?simple" \
  | jq -r '.[] | [.name, .properties.segments.count, .properties.segments.size] | @tsv' | column -t | head -15

echo ""
echo "=== Services ==="
curl -s "$DRUID_API/druid/coordinator/v1/servers?simple" \
  | jq -r '.[] | [.host, .type, .tier, .currSize, .maxSize] | @tsv' | column -t | head -10

echo ""
echo "=== Supervisors ==="
curl -s "$DRUID_API/druid/indexer/v1/supervisor" \
  | jq -r '.[] | [.id, .state // "unknown"] | @tsv' | column -t | head -10

echo ""
echo "=== Load Queue ==="
curl -s "$DRUID_API/druid/coordinator/v1/loadqueue?simple" \
  | jq -r 'to_entries[] | [.key, .value.segmentsToLoad, .value.segmentsToDrop] | @tsv' | column -t | head -10

Analysis Phase

#!/bin/bash

DRUID_API="${DRUID_ROUTER_URL:-http://localhost:8888}"

echo "=== Running Tasks ==="
curl -s "$DRUID_API/druid/indexer/v1/tasks?state=running" \
  | jq -r '.[] | [.id[:40], .type, .dataSource, .createdTime] | @tsv' | column -t | head -10

echo ""
echo "=== Failed Tasks (Recent) ==="
curl -s "$DRUID_API/druid/indexer/v1/tasks?state=failed&max=10" \
  | jq -r '.[] | [.id[:40], .dataSource, .createdTime, .errorMsg[:50] // ""] | @tsv' | column -t

echo ""
echo "=== Segment Distribution ==="
curl -s "$DRUID_API/druid/coordinator/v1/datasources/$DRUID_DATASOURCE/intervals?simple" \
  | jq -r 'to_entries[:10][] | [.key, .value.count, .value.size] | @tsv' | column -t

echo ""
echo "=== Compaction Status ==="
curl -s "$DRUID_API/druid/coordinator/v1/compaction/status" \
  | jq -r '.latestStatus[] | [.dataSource, .byteCountAwaitingCompaction, .segmentCountAwaitingCompaction] | @tsv' | column -t | head -10

echo ""
echo "=== Query Metrics ==="
curl -s -X POST "$DRUID_API/druid/v2/sql" \
  -H "Content-Type: application/json" \
  -d '{"query": "SELECT datasource, COUNT(*) AS num_segments, SUM(size) AS total_size, SUM(num_rows) AS total_rows FROM sys.segments WHERE is_published = 1 GROUP BY datasource ORDER BY total_size DESC LIMIT 10"}' \
  | jq -r '.[] | [.datasource, .num_segments, .total_size, .total_rows] | @tsv' | column -t

echo ""
echo "=== Supervisor Status ==="
for SUP_ID in $(curl -s "$DRUID_API/druid/indexer/v1/supervisor" | jq -r '.[:5][].id'); do
  curl -s "$DRUID_API/druid/indexer/v1/supervisor/$SUP_ID/status" \
    | jq -r '[.id, .payload.state, .payload.healthy] | @tsv'
done | column -t

Output Format

CLUSTER
Version:     <version>
Services:    <count> nodes

DATASOURCES
Name             Segments   Size
<datasource>     <n>        <bytes>

SUPERVISORS
ID               State       Healthy
<supervisor>     RUNNING     true

RUNNING TASKS
ID               Type        Datasource      Created
<task-id>        <type>      <datasource>    <timestamp>

FAILED TASKS
ID               Datasource      Created         Error
<task-id>        <datasource>    <timestamp>     <message>

COMPACTION
Datasource       Bytes Awaiting   Segments Awaiting
<datasource>     <bytes>          <n>

Anti-Hallucination Rules

1. NEVER assume resource names — always discover via CLI/API in Phase 1 before referencing in Phase 2.
2. NEVER fabricate metric names or dimensions — verify against the service documentation or --help output.
3. NEVER mix CLI commands between service versions — confirm which version/API you are targeting.
4. ALWAYS use the discovery → verify → analyze chain — every resource referenced must have been discovered first.
5. ALWAYS handle empty results gracefully — an empty response is valid data, not an error to retry.

Counter-Rationalizations

Shortcut	Counter	Why
"I'll skip discovery and check known resources"	Always run Phase 1 discovery first	Resource names change, new resources appear — assumed names cause errors
"The user only asked for a quick check"	Follow the full discovery → analysis flow	Quick checks miss critical issues; structured analysis catches silent failures
"Default configuration is probably fine"	Audit configuration explicitly	Defaults often leave logging, security, and optimization features disabled
"Metrics aren't needed for this"	Always check relevant metrics when available	API/CLI responses show current state; metrics reveal trends and intermittent issues
"I don't have access to that"	Try the command and report the actual error	Assumed permission failures prevent useful investigation; actual errors are informative