Skip to content

settings.md

Latest commit

 

History

History
55 lines (38 loc) · 2.69 KB

settings.md

File metadata and controls

55 lines (38 loc) · 2.69 KB

Settings

Global settings may be provided at the top level of the config. It is done in form of special commands.

Time settings

When Beholder creates a datetime (e.g. when a new message is received), it uses system timezone. You can override this by setting create_dates_in_timezone to a desired timezone. See java.time.ZoneId.of() docs for details. Example timezones: UTC, Europe/Moscow. When a date is received in a message (not created), Beholder will try to keep its timezone intact.

create_dates_in_timezone UTC;

Buffers and queues

When messages cannot be processed quick enough, they are stored in queues. When the config is being reloaded, incoming messages are stored in from-queues. When a destination (e.g. a TCP server) is unavailable, outgoing payloads are stored in to-queues.

queue_chunk_messages 500;

A queue will group messages into chunks. When there are a lot of chunks in a queue, they get stored in a buffer. Currently there is only one buffer for all this data: the default buffer. Buffer size can be changed using memory_bytes option. You can specify a number of bytes or use a suffix: k, m, g.

buffer {
    memory_bytes 128m;
    memory_compression lz4-fast; # One of: lz4-fast, off
}

The buffer command can only be used at the top level of the config (not inside flow or anywhere else).

When a queue chunk is buffered, it is compressed. You can turn off compression with memory_compression off inside buffer command.

Note: memory_bytes is NOT the total memory limit of Beholder. In addition to unbuffered messages in queues, there are lots of things on JVM that eat bytes. Watch your metrics!

Prometheus metrics

Beholder can report its metrics for Prometheus over HTTP.

prometheus_metrics_http_address 127.0.0.1:8080;

If the setting is specified at top level of the config, Beholder will start a HTTP server on specified address that will respond to GET requests (on any URL path) with a subset of internal Beholder stats in Prometheus text format.

Some stats that are available via parse beholder-stats only make sense if stat data is reset every time the stat is read. For Prometheus, we do not reset stats when they are fetched, an so for example fromTcpMaxBytes stat will not have any useful information, because it will report maximum packet length from the beginning of time. Such stats will not be reported into Prometheus. When metrics are pushed into a database by Beholder itself, it is usually done in regular intervals, and therefore all stats will provide useful information (for example, maximum packet length received in last 30 seconds).