monitoring – Page 32 – Türker Yıldırım

Symfony 4.3 shipped an HTTP client worth using

Deep diveMay 14, 201910 min read

The framework had no client of its own and now it has one with a reason to exist: responses that are lazy by default, and concurrency without a…

Site health, and the plugin that was quietly failing

Deep diveMay 9, 201910 min read

5.2 added a screen that tells you what is wrong, which is new for WordPress. What the checks actually test, and the recovery email nobody receives.

A cache that survives the thing it is caching

Deep diveMay 8, 201914 min read

The cache expired at noon and the database fell over at 12:00:01. A lock, a two-tier expiry, jittered TTLs, and versioned keys — in that order of importance.

Elasticsearch 7 removed mapping types

Deep diveApril 30, 201910 min read

One type per index in 6, none at all in 7, and every query URL changes. The reindex cannot be done in place — and the default shard…

performance_schema knows which indexes are never read

NoteApril 17, 20191 min read

Guessing which of forty indexes on a table are dead is how they stay for years; the server has been counting reads per index the whole time. Counters…

journalctl –output=json and a log pipeline for free

NoteMarch 26, 20191 min read

The journal already stores structured records with unit, priority, hostname and timestamps; the text output is a rendering, not the data. Following the journal in JSON and piping…

The fatal error recovery email goes to an address nobody reads

NoteMarch 7, 20192 min read

5.2 catches a fatal error during a plugin load, deactivates the plugin, and emails a recovery link to the administration email address — which on most sites is…

Beanstalkd bury is a dead letter state built into the server

NoteMarch 4, 20192 min read

Most brokers make a dead letter queue a convention you assemble from routing rules; Beanstalkd has it as one of three verbs. A buried job sits in its…

A for clause on every alert rule, or it fires on noise

NoteMarch 4, 20192 min read

A threshold evaluated instantaneously fires on every transient spike, and transient spikes are constant — which is how a channel accumulates three hundred alerts a month and gets…

Site health checks are a filter, so you can add your own

NoteFebruary 28, 20192 min read

5.2 shipped a screen that reports on the installation, and the useful part is that the list of checks is filterable — so a site can test the…