Agents

Learn about the different types of agents and how to set up your own agent.

Types of Agent

There are two types of agents:

  1. Hosted agents. These are included in your DataBug plan and are operated by us on your behalf.

  2. Private agents. You host these by running our Docker image on your own on-premise or Cloud infrastructure.

Your plan will include a limited number of hosted agent hours per month. As hosted agents carry out the steps described in your Problem Definition, that time will be eaten up on a per-second basis. We do this to ensure that all our customers have fair access to our agent pool.

With private agents, you can use them as much as you like without any quotas since they're running on the infrastructure you're already paying for yourself.

In a future update, we will allow hosted agents to connect to data sources via an SSH tunnel and port mapping.

The concept of "Place"

An agent also runs in a place. A place might be somewhere that's operated by us or one that you own.

DataBug Places

We have a presence in data centres around the world. When running a query, you may decide that the optimum place to run a particular query is in the EU data centre.

This choice might be driven by proximity to the data; an agent close to the data should run more quickly and use less of your monthly agent time. It may also be driven by data protection legislation in your area, such as GDPR.

Here, you'd specify a place of eu and we'd allocate the Problem Definition to an agent running in that data centre.

The available DataBug places are:

  • uk - the United Kingdom

  • us - the United States

  • eu - the European Union

More places will be added soon!

Private Places

When you host your own agent, you will be able to define your own custom places.

Each place name is made up of two parts:

  1. Your customer identifier which is assigned by us (e.g. A5QRY)

  2. A name that you specify.

For example, private place names might include:

  • A5QRY-LosAngeles

  • A5QRY-London

  • A5QRY-Factory

You can then specify where the steps in your Problem Definition are carried out. This would allow you to query one database that's only available within the LA office, and another that's only available in the London office, for example. You could then bring these two datasets together for analysis.

One private place can have multiple agents running. Where that's the case, whichever agent isn't busy will pick up the work. If all agents are busy, their work will stay queued for them.

Deploying a Private Agent

To deploy your own agent, see Deploying your own Agent.

Last updated

Was this helpful?