managing-hazelcast

name	managing-hazelcast
description	Use when working with Hazelcast — hazelcast cluster management, distributed data structure inspection, partition health monitoring, and near-cache analysis. Covers member discovery, map/cache statistics, WAN replication status, CP subsystem health, and executor service diagnostics. Read this skill before any Hazelcast operations.
connection_type	hazelcast
preload	false

Hazelcast Management Skill

Monitor, analyze, and optimize Hazelcast clusters safely.

MANDATORY: Discovery-First Pattern

Always check cluster state and member list before inspecting data structures. Never assume map names or partition counts.

Phase 1: Discovery

#!/bin/bash

HZ_URL="${HAZELCAST_URL:-http://localhost:5701}"

echo "=== Cluster State ==="
curl -s "$HZ_URL/hazelcast/rest/cluster" 2>/dev/null || echo "REST API may be disabled"

echo ""
echo "=== Management Center API ==="
MC_URL="${HAZELCAST_MC_URL:-http://localhost:8080}"
curl -s "$MC_URL/rest/clusters" 2>/dev/null | python3 -c "
import sys, json
data = json.load(sys.stdin)
for cluster in data if isinstance(data, list) else [data]:
    print(f\"Cluster: {cluster.get('name','?')} | Members: {cluster.get('memberCount','?')} | State: {cluster.get('state','?')}\")
" 2>/dev/null || echo "Management Center not available"

echo ""
echo "=== Health Check ==="
curl -s "$HZ_URL/hazelcast/health" 2>/dev/null

echo ""
echo "=== Members via REST ==="
curl -s "$HZ_URL/hazelcast/rest/management/cluster/state" 2>/dev/null

echo ""
echo "=== Using hz-cli if available ==="
if command -v hz-cli &>/dev/null; then
    hz-cli cluster
    hz-cli list-jobs 2>/dev/null || true
else
    echo "hz-cli not found in PATH"
fi

Phase 1 outputs: Cluster state, member count and addresses, health status, Management Center availability.

Phase 2: Analysis

#!/bin/bash

MC_URL="${HAZELCAST_MC_URL:-http://localhost:8080}"
CLUSTER="${1:-dev}"

echo "=== Map Statistics ==="
curl -s "$MC_URL/rest/maps/$CLUSTER" 2>/dev/null | python3 -c "
import sys, json
data = json.load(sys.stdin)
for m in (data if isinstance(data, list) else []):
    print(f\"Map: {m.get('name','?')} | Entries: {m.get('ownedEntryCount',0)} | Memory: {m.get('ownedEntryMemoryCost',0)//1024}KB | Hits: {m.get('hits',0)}\")
" 2>/dev/null | head -15

echo ""
echo "=== Partition Distribution ==="
curl -s "$MC_URL/rest/clusters/$CLUSTER/partitions" 2>/dev/null | python3 -c "
import sys, json
data = json.load(sys.stdin)
member_counts = {}
for p in (data if isinstance(data, list) else []):
    owner = p.get('owner','?')
    member_counts[owner] = member_counts.get(owner, 0) + 1
for member, count in sorted(member_counts.items()):
    print(f\"  {member}: {count} partitions\")
" 2>/dev/null || echo "Partition info not available"

echo ""
echo "=== WAN Replication ==="
curl -s "$MC_URL/rest/wan/$CLUSTER" 2>/dev/null | head -10 || echo "WAN replication not configured"

echo ""
echo "=== CP Subsystem ==="
curl -s "$HZ_URL/hazelcast/rest/cp-subsystem/groups" 2>/dev/null || echo "CP subsystem not enabled"

Output Format

HAZELCAST ANALYSIS
==================
Cluster: [name] | State: [active/frozen/passive]
Members: [count] | Partitions: [count]

ISSUES FOUND:
- [issue with affected member/map]

RECOMMENDATIONS:
- [actionable recommendation]

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