Skip to main content

Tecton Support

Support tickets can be created using the following methods:

Tecton’s support portal at support.tecton.ai is preferred because:

  • The priority of the ticket can be specified (see below)
  • Tickets created through the portal are directly routed to the relevant teams internally to ensure SLA guarantees are met (SLAs are only measured for tickets logged in the portal)
  • The portal can be used to view previous support tickets submitted by you or your team.

Prior to creating a ticket, we recommend reviewing the Tecton documentation for assistance.

Creating a support ticket via support.tecton.ai​

  • Prerequisite (one-time process): To create a support ticket, you must be added to the support portal. To request access, please contact your Account Representative or the Support Team at techsupport@tecton.ai.
  • Navigate to support.tecton.ai and log in using the same email and SSO method used for your Tecton deployment.

Creating a support ticket via email​

Send an email to techsupport@tecton.ai with a summary and description of your issue.

Priority​

The following table explains what priority level you should use when filing a tickets. Priority can only be specified for tickets created using support.tecton.ai.

Priority LevelWhen to use
P0Critical: Due to a Tecton problem, my production model is down, causing a major outage. I have no workarounds. This priority will page our on-call engineer 24/7. Please do not use for feature services not in active production use.
P1High: I am using Tecton in production, and due to a Tecton problem, my production model is seriously degraded. I may have a workaround, but it is cumbersome. This priority level is also reserved for customers in production.
P2Medium: For a question or technical support issue in either production or development, and I am delayed while getting resolution to this question.
P3Low: For a question or technical support issue that is not impacting production serving or blocking development.

What information should I provide to Tecton to troubleshoot my issue?​

Please provide the following information, according to the issue type.

Issue typeWhat we need
Notebook-related issues, including slow/failing/incorrect Offline Retrieval
  • (Check our Offline Retrieval troubleshooting support articles first!)
  • ipynb notebook containing a concise but full reproduction of the issue, including any Tecton stack trace
  • Zip file with your feature repository, or cluster and workspace name.
  • Link to the EMR or Databricks notebook cluster id you ran this command on
  • Tecton SDK version on your notebook cluster (tecton.version.summary())
  • How long did the GHF command take to run before it failed?
  • What you did to try to fix the problem (briefly)
  • Ideal: “Debugging” access so we can run notebooks in your cluster. This will significantly decrease the time it takes to resolve your issue. Otherwise, we will need to debug this asynchronously with you. Contact Tecton support for more information on this.
    • HTTP API issues
  • Cluster name, workspace name, and feature service name
  • The HTTP response code(s)
  • Error message that Tecton returns
  • JSON request payload (if applicable)
    • Tecton plan/apply errors
  • Zip file with feature repository you’re attempting to plan or apply
  • Full output from your tecton plan/apply, including full error message
  • Cluster name and workspace name you’re trying to plan/apply to
  • Tecton CLI version (tecton version)
    • General feature implementation questionsConcise repro of the feature code you have a question on that can be plan/apply’d to a workspace
    Materialization problems
  • If spot instance failures, see this support article.
  • Link to the feature view in your cluster’s web UI whose materialization is failing
  • Link to an example materialization job cluster that has failed
    • Infrastructure requests
  • What specifically you would like Tecton to do (increase qps capacity, etc.)
  • Time constraints: when should Tecton start the work? Any blackout times Tecton should honor?