What do I do if Kosli is down?

This page shows you how to bypass Kosli attestations if Kosli is down so your CI pipelines keep running, and how to re-enable them when it recovers.

Status of Kosli services

You can also check the current status of Kosli services by running kosli status in the CLI.

Turning Kosli CLI calls on and off instantly

If the KOSLI_DRY_RUN environment variable is set to true then all Kosli CLI commands will:

• Not communicate with Kosli at all
• Print the payload they would have sent
• Exit with a zero status code

We recommend creating an Org-level KOSLI_DRY_RUN variable in your CI system and, in all CI workflows, ensuring there is an environment variable set from it.

For example, in a Github Action workflow:

name: Main
...
env:
  KOSLI_DRY_RUN: ${{ vars.KOSLI_DRY_RUN }}           # true iff Kosli is down

Turning Kosli API calls on and off instantly

If you are using the Kosli API in your workflows (e.g. using curl), we recommend using the same Org-level KOSLI_DRY_RUN environment variable and guarding the curl call with a simple if statement. For example:

#!/usr/bin/env bash

kosli_curl()
{
  local URL="${1}"
  local JSON_PAYLOAD="${2}"

  if [ "${KOSLI_DRY_RUN:-}" == "true" ]; then
    echo KOSLI_DRY_RUN is set to true. This is the payload that would have been sent
    echo "${JSON_PAYLOAD}" | jq .
  else
    curl ... --data="${JSON_PAYLOAD}" "${URL}"
  fi
}