8 Commits (v3.0_fix_codecov-coverage-path-prefix)

Author	SHA1	Message	Date
Claude Code	e548f749b0	monitor: drop unreachable DEFAULT on port column in admin SQLite tables ... The 'port' column in 17 active MONITOR_SQLITE_TABLE_* macros carried 'DEFAULT 3306' (or 'DEFAULT 6032' for proxysql_servers). That default is unreachable: every table's PRIMARY KEY contains 'port NOT NULL', and all internal INSERTs from the monitor supply a port explicitly. The default is only visible via 'DESCRIBE' — it never affects a row, a query, or a runtime path. Drop the clause: - 5 macros in include/PgSQL_Monitor.hpp (5x DEFAULT 3306, 1x DEFAULT 6032) - 11 active macros in include/MySQL_Monitor.hpp (10x DEFAULT 3306, 1x DEFAULT 6032) The 5 '//#define' commented-out lines in MySQL_Monitor.hpp are left untouched as dead-code historical reference. No migration concern: SQLite's 'CREATE TABLE IF NOT EXISTS' won't recreate existing admin databases, but since the dropped defaults were never reached, no data, query, or behavior changes either way. Side note: the 5 MONITOR_SQLITE_TABLE_PGSQL_* macros in include/PgSQL_Monitor.hpp are not referenced from any .cpp file under lib/ or src/ — they appear to be scaffolded for future PgSQL monitor work. Out of scope here. Build verified: 'make debug' (O0 -ggdb -DDEBUG) compiles and links the proxysql binary cleanly with the changes.	2 weeks ago
Rene Cannao	031f6fa1f5	fix(build): adapt callers of dns_cache counters to std::atomic ... The previous commit promoted MySQL_Monitor::dns_cache_{queried, lookup_success,record_updated} to std::atomic<unsigned long long> but missed several call sites that pass them through implicit copy: * lib/MySQL_Thread.cpp sprintf("%llu", GloMyMon->dns_cache_*) — varargs cannot copy std::atomic. * lib/PgSQL_Thread.cpp sprintf("%llu", GloPgMon->dns_cache_*) — same, for the PgSQL counters. * lib/MySQL_Monitor.cpp p_update_counter(..., GloMyMon->dns_cache_*) — works via implicit operator T(), but explicit .load() is clearer and matches the rest of the file. Promote PgSQL_Monitor::dns_cache_* to std::atomic for the same race that motivated the MySQL side (resolver workers increment, stats readers and p_update_metrics read from another thread), and add .load() at every read site so we don't depend on implicit conversion. Verified with a local 'make -j32 debug' build in the integration worktree.	1 month ago
Rene Cannao	ca145d5026	fix(pgsql): atomic shutdown flag; honor dns_lookup fallback contract ... GloPgMon->shutdown was a plain bool, written by main.cpp's teardown and read by the monitor / DNS resolver loops on worker threads — a data race that can let the loops spin past shutdown or observe a stale value indefinitely. Promote to std::atomic<bool> and switch the readers/writer to acquire/release ordering. PgSQL_Monitor::dns_lookup() was also returning an empty string when GloPgMon / dns_cache_thread weren't initialized yet, ignoring the return_hostname_if_lookup_fails contract. Move the fallback out of the dns_cache_thread branch so startup / shutdown behave like a normal cache miss for callers that asked for the hostname fallback. Addresses CodeRabbit review feedback on PR #5806.	1 month ago
Rene Cannao	a19894427e	Add independent DNS cache for PgSQL ... Fixes the watchdog asserts under DNS degradation reported in #5768. Symptom: with a PgSQL backend addressed by hostname, an unhealthy resolver made every PQconnectStart() block ~5s on getaddrinfo inside libpq. Under load, all PgSQL_Thread workers spent the bulk of each loop iteration in that synchronous DNS call, atomic_curtime fell behind the watchdog threshold (poll_timeout + 1s), and after restart_on_missing_heartbeats misses the watchdog asserts. MySQL avoids this because mysql-monitor's DNS cache lets MySQL_Connection::connect_start pass an IP directly to libmariadb; PgSQL had no equivalent. This change adds the equivalent for PgSQL, with state fully independent of the MySQL cache: * PgSQL_Monitor gains its own DNS_Cache, force_dns_cache_update flag, and dns_cache_{queried,lookup_success,record_updated} counters. PgSQL_Monitor::dns_lookup / update_dns_cache_from_pgsql_conn / trigger_dns_cache_update mirror the MySQL_Monitor surface. * A new background thread (PgSQL_monitor_dns_cache_pthread) drives PgSQL_Monitor::monitor_dns_cache(), which walks PgHGM->pgsql_servers_to_monitor and feeds the shared DNSResolverWorker pool. Launched alongside PgSQL_monitor_scheduler and lives independently of pgsql-monitor_enabled, matching MySQL. * PgSQL_Connection::connect_start now consults the cache. On a hit it appends `hostaddr=<ip>` to the libpq conninfo while keeping `host=<hostname>`; libpq documents this combo specifically to skip name resolution while preserving TLS hostname verification. On a miss / IP literal / disabled cache we fall back to the previous behavior (libpq does getaddrinfo). * Successful PgSQL connects seed the cache via update_dns_cache_from_pgsql_conn, so the second connect avoids getaddrinfo even before the resolver loop has visited the host. * The pgsql-monitor_local_dns_cache_ttl / _refresh_interval / _resolver_queue_maxsize variables (already accepted by Admin via PgSQL_Thread's variable registry) now actually propagate to pgsql_thread___monitor_local_dns_* globals. The propagation block in PgSQL_Thread::refresh_variables was previously commented out, so the values never reached anywhere. * The runtime stats output (SQL3_GlobalStatus) now exposes the PgSQL_Monitor_dns_cache_* counters alongside the existing PgSQL_Monitor_* counters. DNSResolverWorker is shared infrastructure added to DNS_Cache.hpp/cpp, used here by PgSQL_Monitor. MySQL_Monitor still uses its existing ConsumerThread-based pool; nothing about the MySQL cache changes.	1 month ago
Javier Jaramago Fernández	ba5af979ff	feat-wip: Add replication lag support for PostgreSQL monitoring	4 months ago
Rahim Kanji	fae283cf7e	Add SSL and non-SSL connection OK metrics for PostgreSQL monitor connections ... Adds two new metrics, ssl_connections_OK and non_ssl_connections_OK, to improve visibility into PostgreSQL monitor connection status.	7 months ago
Javier Jaramago Fernández	100630fba5	Second iteration on PostgreSQL monitoring POC ... This commit packs multiple fixes/improvements: - Added READONLY support for PostgreSQL. - Major rework for queries and statements used on monitoring checks: + Checks/Actions rewrote for single instance checks. + Reuse of prepared statements instance of re-preparation. - Fixed missing error handling in connection creation state machine. - Fixed connection rotation in connection pool (now FIFO). - Added support for configurable batching in scheduler thread, via '*_interval_window' variables. These variables allows to define the burstiness of the scheduling within the processing interval. - Added new config variable 'pgsql-monitor_dbname'. Allows to control which 'db' will be target by monitoring connections. - Several fixes for 'poll' timeout computation for worker threads. - Fixed edge cases for current interval detection. - Reduced deviation in scheduling intervals computation. - Refactored and simplified connection event handling. - Improved error messages for monitoring actions. - Replaced several invalid uses of 'mysql_thread___monitor_*' in favor or new 'mysql_thread___monitor_*' variables. - Honor '-M' argument for disabling monitoring support.	2 years ago
Javier Jaramago Fernández	705f50e0f9	Add initial POC for PostgreSQL monitoring support	2 years ago