An HTTP sidecar service, wrapping the great open source bull library, for easy to use job scheduling. Bull is stable, well docummented library, that uses Redis as an underlying DB to manage jobs and queues. The sidecar behaved as both the producer and the consumer of the defined job type/queue. Exposes prometheus metrics for monitoring and alerts

• Need X to happen Y minutes from now
• Possibly long periods of time- an hour or more

Example:

Say we have a chat service, where, when a user receives a message, he needs to be reminded or notified to read it after 1 minute, if he hasn't done so yet. So we have a service, which listens to some kind of message queue, and upon each arriving message, it schedules a job to be executed in 1 minute. The job is scheduled via sending a request to the sidecar's POST endpoint, and scheduling a job with all the necesary data in it (the recepient's and sender's id and name). When the job's turn arrives, the sidecar sends a POST request to a predefined endpoint on the service, with all the job's data (that was send to the sidecar), and then the service can check if the user has already seen the message, and notify him if not.

The docker image is hosted on docker hub and can be found here: https://hub.docker.com/r/mishaled/postponed/

An example kubernetes replication controller can be found in the examples directory.

Exposes a simple api

• Schedule a Job

  • https://localhost:**HTTP_PORT**

    • POST

    • Will call the host back when the job is active on TARGET_BASE_PATH + TARGET_PROCESS_RELATIVE_PATH via a post request with the jobData as the body

    • Expected body

      {
          jobId?: string;
          jobData: Record<string, any>;
          delayInSec: number;
      }

• Remove a Job

  • https://localhost:**HTTP_PORT**/:id
    • DELETE
    • id is the id of the job we need to remove