managing-yugabytedb

name	managing-yugabytedb
description	Use when working with Yugabytedb — yugabyteDB cluster management, tablet distribution analysis, YSQL/YCQL performance tuning, and replication monitoring. Covers master/tserver health, tablet leader balancing, xCluster replication, CDC streams, and distributed transaction diagnostics. Read this skill before any YugabyteDB operations.
connection_type	yugabytedb
preload	false

YugabyteDB Management Skill

Monitor, analyze, and optimize YugabyteDB clusters safely.

MANDATORY: Discovery-First Pattern

Always check cluster health and list databases/keyspaces before any query operations. Never assume table names or tablet counts.

Phase 1: Discovery

#!/bin/bash

YB_MASTER="${YUGABYTE_MASTER:-localhost:7100}"
YB_TSERVER="${YUGABYTE_TSERVER:-localhost:9000}"
YB_HOST="${YUGABYTE_HOST:-localhost}"

echo "=== Master Health ==="
curl -s "http://${YB_MASTER%:*}:7000/api/v1/health-check" 2>/dev/null || \
    curl -s "http://${YB_MASTER%:*}:7000/status" 2>/dev/null

echo ""
echo "=== Cluster Config ==="
curl -s "http://${YB_MASTER%:*}:7000/api/v1/cluster-config" 2>/dev/null | python3 -c "
import sys, json
data = json.load(sys.stdin)
print(f\"Replication factor: {data.get('replication_info',{}).get('replication_factor','?')}\")
print(f\"Cluster UUID: {data.get('cluster_uuid','?')}\")
" 2>/dev/null

echo ""
echo "=== Tablet Servers ==="
curl -s "http://${YB_MASTER%:*}:7000/api/v1/tablet-servers" 2>/dev/null | python3 -c "
import sys, json
data = json.load(sys.stdin)
for ts in data.get('servers', data if isinstance(data, list) else []):
    print(f\"TServer: {ts.get('host','?')} | Alive: {ts.get('alive','?')} | Tablets: {ts.get('tablet_count','?')} | Read ops: {ts.get('read_ops_per_sec','?')}\")
" 2>/dev/null

echo ""
echo "=== YSQL Databases ==="
ysqlsh -h "$YB_HOST" -c "SELECT datname, pg_size_pretty(pg_database_size(datname)) as size FROM pg_database WHERE datistemplate = false;" 2>/dev/null || \
    psql -h "$YB_HOST" -p 5433 -U yugabyte -c "SELECT datname, pg_size_pretty(pg_database_size(datname)) as size FROM pg_database WHERE datistemplate = false;" 2>/dev/null

echo ""
echo "=== YCQL Keyspaces ==="
ycqlsh "$YB_HOST" -e "DESCRIBE KEYSPACES;" 2>/dev/null || echo "YCQL not available"

Phase 1 outputs: Master health, tablet server list, database/keyspace inventory, replication factor.

Phase 2: Analysis

#!/bin/bash

YB_MASTER="${YUGABYTE_MASTER:-localhost:7100}"
YB_HOST="${YUGABYTE_HOST:-localhost}"

echo "=== Tablet Distribution ==="
curl -s "http://${YB_MASTER%:*}:7000/api/v1/tablet-servers" 2>/dev/null | python3 -c "
import sys, json
data = json.load(sys.stdin)
for ts in data.get('servers', data if isinstance(data, list) else []):
    leader = ts.get('leader_count', '?')
    total = ts.get('tablet_count', '?')
    print(f\"  {ts.get('host','?')}: leaders={leader} total={total}\")
" 2>/dev/null

echo ""
echo "=== Under-Replicated Tablets ==="
curl -s "http://${YB_MASTER%:*}:7000/api/v1/tablets?state=UNDER_REPLICATED" 2>/dev/null | head -10

echo ""
echo "=== YSQL Active Queries ==="
psql -h "$YB_HOST" -p 5433 -U yugabyte -c "SELECT pid, state, query_start, query FROM pg_stat_activity WHERE state != 'idle' LIMIT 10;" 2>/dev/null

echo ""
echo "=== Table Sizes ==="
psql -h "$YB_HOST" -p 5433 -U yugabyte -c "SELECT schemaname, tablename, pg_size_pretty(pg_total_relation_size(schemaname||'.'||tablename)) as size FROM pg_tables WHERE schemaname NOT IN ('pg_catalog','information_schema') ORDER BY pg_total_relation_size(schemaname||'.'||tablename) DESC LIMIT 15;" 2>/dev/null

Output Format

YUGABYTEDB ANALYSIS
===================
Cluster: [uuid] | RF: [factor] | TServers: [count]
Databases (YSQL): [count] | Keyspaces (YCQL): [count]

ISSUES FOUND:
- [issue with affected table/tablet]

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