When you open a support case for Hybrid Manager (HM) or for a Postgres cluster that HM manages, EDB Support might ask you to provide diagnostic data. EDB provides two tools to collect it: Lasso and the EDB Diagnostic kubectl plugin.
Lasso
Lasso is a diagnostics application from EDB. It gathers configuration, statistics, and system information from a Postgres server and its environment. Lasso doesn't collect the data stored in your user tables. You can inspect the files it gathers before you send them to EDB.
For HM, Lasso is available as a container image in the pgai-platform repository on the EDB container registry. The EDB Diagnostic plugin uses this image to collect data from a Postgres cluster. You can also run Lasso directly on a host where Postgres runs.
- For the full list of data that Lasso collects, see Lasso collected data.
- To run Lasso on a host without Kubernetes, see the knowledge base article (EDB account required) Running Lasso to send diagnostic information.
EDB Diagnostic kubectl plugin
The EDB Diagnostic plugin (kubectl-edb-diagnostic) orchestrates diagnostic collection on a Kubernetes cluster. You install it on your own workstation. It connects to the Kubernetes API using your existing role-based access control (RBAC) permissions, so you don't need to go through your organization's container image onboarding process to install the plugin itself. In database mode, the plugin starts a pod that runs the Lasso image, so your Kubernetes cluster must be able to pull that image. By default, the plugin uses the Lasso image from the EDB container registry. You can specify a different image with the --imagename option.
The plugin runs in one of two modes:
| Mode | Use it when | What it does |
|---|---|---|
| Database | You have a problem with a specific Postgres cluster, such as degraded performance. | Targets the namespace of the Postgres cluster. The plugin starts a pod that runs the Lasso image, and Lasso connects to the database and collects data. |
| Service | You have a general installation or infrastructure problem, such as a component that failed or a console that doesn't load. | Collects Kubernetes information across the cluster, including namespaces, pods, descriptions, and logs. This mode doesn't use Lasso. It's also the best way to collect data when a Postgres cluster fails because of a misconfigured resource, for example, a wrong secret reference. |
In both modes, the plugin writes the output as compressed files in the directory where you ran it, or in the directory you specify. Service mode produces one file for the cluster. Database mode produces one file for each database pod in the Postgres cluster. Send the files to EDB Support by attaching them to your support case.
For installation and usage instructions, see the knowledge base article (EDB account required) How to use kubectl-edb-diagnostic to collect Hybrid Manager cluster diagnostics.