Free audit

View Audit Scope

Elasticsearch incident support

JusDB provides Elasticsearch incident support for a cluster that is yellow or red, for shards that will not assign, and for a circuit breaker that is rejecting requests. Your team keeps day-to-day operations. JusDB joins the incident under a service agreement. This page is not a design review. Start by sending the incident details.

When do teams ask for Elasticsearch support?

Teams ask when the cluster is yellow or red, a shard will not assign, or a circuit breaker is rejecting requests.

  • The cluster is yellow or red

    Yellow means every primary is assigned and at least one replica is not. Red means at least one primary is unassigned. That is Elastic's health model.

  • A shard will not assign

    Disk watermarks, a node that left, or a version mismatch can leave a shard unassigned. The allocation explain API names the decision.

  • A breaker is rejecting requests

    Parent, request, fielddata, and in-flight breakers protect the heap. JusDB reads the breaker stats before changing a limit.

  • The design is already settled

    Index design and capacity belong on the consulting page. Support joins the incident your team cannot close.

What does Elasticsearch support cover?

Support covers health and allocation, breaker pressure, and the incident window. Design stays on the consulting page.

  • Health and allocation

    JusDB reads cluster health and the allocation explain output before moving a shard or changing a watermark.

  • Breaker pressure

    JusDB reads which breaker tripped and which request caused it, then proposes a limit or a query change.

  • The incident window

    A node restart or a setting change can interrupt queries. The window and the rollback are agreed before the edit.

  • Hand-back

    Your team keeps day-to-day operations. JusDB leaves a note of the signal to watch after the incident.

How does an Elasticsearch incident run?

JusDB reads the health and allocation output, agrees the edit, applies that edit, and checks the same signals again.

  1. 1

    Send the incident

    Send the health color, the error, and when it started.

  2. 2

    Agree access

    Work starts once scope and access are agreed. A retainer that already has read-only access skips that wait.

  3. 3

    Read the explain output

    JusDB reads allocation explain, cluster health, and breaker stats before changing a setting.

  4. 4

    Apply the agreed edit

    JusDB applies only the edit named for this incident.

  5. 5

    Confirm the shard state

    JusDB checks the same health and allocation signals after the edit.

What does this page not claim?

This page states the limit of the offer. Results depend on the workload; we baseline before and measure after each approved change.

  • This page describes incident support. It does not report a client result or a recovery time.
  • Support is reactive. Continuous operations are a separate agreement.
  • A restart or a reroute can interrupt queries. The window is agreed before the edit.
  • Elastic's health colors are Elastic's definitions. JusDB does not redefine them.

Where is the mechanism documented?

Cluster health, shard allocation, and circuit breakers are Elasticsearch mechanisms. JusDB follows Elastic's documentation for the version in production. Official documentation: Elasticsearch.

Questions before you engage

The questions cover scope, how the work starts, the initial-response target, and billing.

How is Elasticsearch support different from consulting?

Consulting is a design review: shards, index lifecycle, and capacity. Support is reactive. Your team keeps day-to-day operations, and JusDB joins when the cluster is yellow or red, a shard will not assign, or a breaker is rejecting requests.

Does Elasticsearch support include a version upgrade?

A version upgrade is planned on the Elasticsearch migration page. Support can stabilize the cluster that is up now. It does not run a major upgrade inside the incident.

How does the engagement start?

If you are not yet a client, send the incident details. We review enquiries within 2–4 business hours, Mon–Fri 09:00–18:00 IST. Work starts once scope and access are agreed.

What initial-response target applies?

Initial-response targets are 2 to 4 hours during covered business hours, Mon–Fri 09:00–18:00 IST (UTC+5:30). Where your service agreement includes 24x7 on-call, the initial-response target for P1 incidents is 30 minutes. Targets are not resolution guarantees.

How is the work billed?

A monthly retainer for a set number of engineering hours, or a monthly fee per database instance, HA cluster, or region; projects are scoped and quoted separately.

Where the neighboring Elasticsearch work lives

Design, upgrades, and Kubernetes each have their own page.

Start with a scoping call

The scoping call agrees the estate and the access. A free audit covers one database instance. After a mutual NDA, the review is least-privilege and read-only. The offer is subject to availability.

Related Elasticsearch Services

Explore more ways our Elasticsearch experts can help with your database infrastructure.