Skip to main content

service logs

To print the logs for services in an enclave, run:

kurtosis service logs $THE_ENCLAVE_IDENTIFIER $THE_SERVICE_IDENTIFIER1 $THE_SERVICE_IDENTIFIER2 $THE_SERVICE_IDENTIFIER3

where $THE_ENCLAVE_IDENTIFIER and the $THE_SERVICE_IDENTIFIER are resource identifiers for the enclave and services, respectively. The service identifier (name or UUID) is printed upon inspecting an enclave. :::

Number of log lines

By default, logs printed in the terminal from this command are truncated at the most recent 200 log lines. For a stream of logs, we recommend the -f flag. For all the logs use the -a flag and for a snapshot of the logs at a given point in time (e.g. after a change), we recommend the kurtosis dump.

Kurtosis will keep logs for up to 4 weeks before removing them to prevent logs from taking up to much storage. If you'd like to remove logs before the retention period, kurtosis enclave rm will remove any logs associated for service in the enclave and kurtosis clean will remove logs for all services in stopped enclaves. :::

The following optional arguments can be used:

  1. -a, --all can be used to retrieve all logs.
  2. -n, --num=uint32 can be used to retrieve X last log lines. (eg. -n 10 will retrieve last 10 log lines, similar to tail -n 10)
  3. -f, --follow can be added to continue following the logs, similar to tail -f.
  4. -x, --all-services can be used to retrieve logs for all services in an enclave. Another option is to pass in the escaped wildcard operator like so kurtosis service logs enclave-name '*'
  5. --match=text can be used for filtering the log lines containing the text.
  6. --regex-match="regex" can be used for filtering the log lines containing the regex. This filter will also work for text but will have degraded performance.
  7. -v, --invert-match can be used to invert the filter condition specified by either --match or --regex-match. Log lines NOT containing the match will be returned.

Important: --match and --regex-match flags cannot be used at the same time. You should either use one or the other.