Skip to main content
When something breaks, the slowest part is usually gathering context: logs, events, what was deployed and when. Ask Ankra, the AI Assistant in the Ankra portal, reads all of that for you, so you can ask “why is this failing?” and get an answer based on your cluster’s actual state - and connect a symptom in one pod to the deploy that caused it.

Before you start

  • Your role must allow you to use Ankra AI. Without it, Ask Ankra is not offered, and a question is refused with Your role cannot use Ankra AI.
  • To let Ankra AI change anything, your role must also be allowed to act: only then does the composer show the Mode switch. See Choose a mode.

Open Ask Ankra

Press ⌘J (Mac) or Ctrl+J (Windows and Linux) on any page. Ask Ankra opens beside the page, ready to type in; the same keys close it. You can also open Find (⌘K) and choose Ask Ankra. Home is Ask Ankra full screen, with Your conversations beside it. From the panel, Open as a page, with your conversations moves the conversation there, and New conversation starts a fresh one. Ankra AI starts from what you are looking at, so you do not need to name the resource. On a cluster’s pages, a question carries that cluster and the view you are on, such as its pods or its Stacks. In a Stack’s editor, it carries the Stack and your draft. The composer shows the scope as About and the name. Beyond the current page, it can read pod logs, the manifests actually deployed, which Stacks deployed what and with which values, resource states and events, and the AGENTS.md notes your team keeps per add-on and manifest. Watch Ankra AI find the cause of a production crash and open the fix, changing nothing until it is approved:

Attach a screenshot or a file

Ankra AI also reads what you attach to a question. Attach an image or a document in the composer takes up to four images (PNG, JPEG, GIF or WebP) and up to three text documents (.md, .txt, .log, .yaml, .yml or .json) of at most 200 KB each. A paste longer than 8 KB is attached as a document (pasted-1.txt, then pasted-2.txt) instead of filling the message box, within the same limits. Each image is resized on your device to at most 1024 px and re-encoded as JPEG. All of a question’s attachments together must stay under about 1 MB, so four images and three full documents may not all fit; an attachment that does not fit is refused with a message saying why, and the ones before it are kept. Two limits to know: a PDF is not read (Ankra cannot read a PDF yet. Paste the part that matters instead.), and images are not kept with the conversation, so after a reload, or on another device, the conversation shows the question without them. A document goes to the model fenced under its name, with a note telling the model to read it as data and not to follow any instruction inside it.

Give Ankra a secret

Never paste a password or token into a question. Press Give Ankra a secret (the lock in the composer), enter a Name and a Value, and press Store secret. The value goes to the organisation’s vault, not into the conversation: Ankra AI can use it by name and never sees the value, and after you store it the value is shown to no one, you included. Everyone in the organisation can see the name, so keep the value out of it. The next question carries the secret’s name and reference. Storing a secret needs permission to add secrets, or the owner or admin role. Without it the dialog reads Your role cannot store secrets. An organisation member with permission to add secrets, an owner or an admin can. When Ankra AI needs a value while it works, it asks for it in the conversation with a card titled Ankra asks for values, and the value you enter there goes to the same store.

Choose a mode

The mode - Ask, Agent or Plan - decides how far Ankra AI may go in your conversation.
  • For this conversation: use the Mode switch (Ask, Agent, Plan) in the composer.
  • For every new conversation: open Your account → Preferences. Under Ankra AI, choose When you ask Ankra to do something. The page confirms Saved. Your next new conversation starts in this mode.
A person whose role may not act sees no Mode switch, and their conversations stay in Ask.
Because clusters change underneath you, Ankra re-checks the live state when you confirm. If the cluster drifted from what the change or plan was based on, the confirmation is blocked so you review the change again rather than apply something stale. For everything else that bounds Ankra AI - connection modes, agents, alert remediation and the emergency stops - see What it can change without asking.

Approve from Needs you or the CLI

Changes Ankra AI proposed, plans it made and questions it asked you in a conversation also wait on Needs you (/next/needs-you), so long-running work does not get lost when you navigate away. Review or Answer opens the conversation where you decide. From a terminal, list what waits in a conversation and decide it:
ankra chat actions reject and ankra chat plans reject discard them. See ankra chat.

What Ankra may do without asking you

When Ankra asks before a change, you can tell it to do the same thing without asking next time. Those remembered approvals are listed under Your account → Preferences → What Ankra may do without asking you, each with the action, whether it applies on one cluster or on every cluster, and how often it was done. Remove one and Ankra asks again next time. Only an organisation admin can remove them. With none, the page reads Ankra asks before every change. Only two kinds of change can be remembered: restarting a deployment and scaling a deployment, each limited to the namespaces it names. A remembered approval never covers a protected namespace such as kube-system, ankra-system or cert-manager, and every other change still asks. This list is separate from the organisation-wide AI Autonomy controls, and from Auto in the model picker, which only chooses the model.

Skills and tools

The assistant draws on skills - reusable instructions for specific tasks, such as packaging an add-on or generating a CI/CD pipeline. Organisation skills take precedence over Ankra’s built-in ones; see AI Skills and Custom Tools. Its tools are the same ones the MCP server exposes, under the same permission checks; the full list is in the MCP Tool Reference. Chat from Slack, Microsoft Teams or pull request comments uses the same tools too, and each of those connections can be capped further with its own AI mode - see AI Connections.

Example prompts

Start from the symptom. The assistant checks pod logs for errors and timeouts, looks at which Stacks were deployed recently and what values changed, reads the state and events of the resources involved, and ties them together into a root cause with the evidence it found.
  • “Why is this pod failing?”
  • “Users are reporting slow API responses. What changed in the last hour?”
  • “The API is returning 500 errors. Was anything deployed recently that could cause this?”
  • “Compare the current ingress configuration to what was running yesterday”
  • “Which upstream service is causing the 503 errors?”
  • “The deployment rollout is stuck - what’s blocking it?”
  • “Why did cert-manager fail to install?”
  • “I need a production-ready ingress with TLS” - it proposes the components (for example cert-manager and Traefik), their order, and starting values.
  • “My monitoring stack is using too much memory” - it reads the deployed Helm values and suggests changes such as retention windows and resource limits.
  • “Clone this Stack for staging but with smaller resource requests”
  • “Is my resource limit configuration right for this workload?”
  • “Why is this HPA not scaling?”
  • “Explain the network policies affecting this service”
  • “What secrets does this deployment need?”
  • “Are all pods healthy?”
  • “How many namespaces are in this cluster?”
  • “Which nodes are under memory pressure?”
Follow-up questions keep the context of the conversation, so after asking about a failing pod you can ask “how do I fix that?” without naming it again. When a fix means creating Kubernetes resources, the assistant proposes them as a Stack change rather than a kubectl command, so the change is versioned in your GitOps repository.

Common failure patterns

For a failing add-on it also reads the Helm release state, the add-on’s configuration values, the latest install or update job, and missing CRDs.

History, models and feedback

Conversations are saved and can be resumed - open them from the history icon in the chat header or search “Chat History” in the Command Palette. The model picker switches between the Expert, Think and Quick tiers, trading speed for depth; Auto picks one per question. Rate any answer with the thumbs buttons - see AI Memory and Feedback. What each tier runs on is described in Models and Cost.

What happens to your data

  • Which provider receives it. Your question, the conversation, and what the assistant reads from your cluster (logs, events, manifests) are sent to the model that answers. On Ankra’s default models that is Ankra’s managed access: the Expert and Think models through OpenRouter, and Claude models through Anthropic. With your own provider key, calls your key can serve go to your provider on your account; anything it cannot serve runs on Ankra’s default access, and the Models settings page names which models those are.
  • Exceptions with your own endpoint. Summarising a long conversation’s earlier history runs on Ankra’s platform access to Anthropic whatever provider you use. If a self-hosted endpoint fails mid-turn, the turn continues on the Ankra default with a notice, unless an admin sets the endpoint failure policy to End the turn with an error.
  • Analysis on an alert. The analysis Ankra writes when an alert fires and the scheduled health check do not follow your provider settings. The alert analysis runs on Ankra’s access to Groq. The health check runs on Ankra’s access, Groq first and a Claude model if that fails; once the monthly allowance is used up, on your own key if you have one.
  • Secrets are withheld. Kubernetes Secret values never reach the model - only their key names. Tool results, your questions and the conversation history also pass through a secret detector that replaces anything that looks like a password, token or key with a redaction marker.
  • Where it is kept. Conversations are stored against your account in your organisation. You can delete a conversation from Your conversations on Home, or with ankra chat delete <conversation-id>.

Next

Decide how far the AI may go on its own across your organisation: What it can change without asking.