Grafana Metrics Migration

Overview

A fixed-price, 3-calendar-week professional services engagement that migrates a representative Grafana metrics estate to Elastic Observability using an ingest-led approach (Path 1). Prometheus metrics land in Elastic first, then the Observability Migration Platform (OMP) converts in-scope Grafana dashboards and selected alert rules against real data. Scope is expressed in countable units — not consulting days — so boundaries are legible to both parties.

Path 1 ingest-led approach: metrics ingested into Elastic with PromQL-compatible shape using Prometheus remote-write, EDOT, or Elastic Agent integrations → OMP conversion applied to the live Elastic project → assisted panel polish within the agreed allowance → alert rule evaluation within the agreed allowance → side-by-side walkthrough.

Scope

• Phase 1: Start-gate verification; project kickoff; ingest approach agreed; alert shortlist and time-box confirmed; post-kickoff confirmation issued
• Phase 2: OMP scorecard review with Customer; datasource inventory; discovery sessions (metric collection architecture, named systems, network constraints, alerting ownership); Path 1 ingest plan, system-to-config mapping, and alert shortlist locked in writing
• Phase 3: Prometheus remote-write and/or up to 6 EDOT/Elastic Agent sample configurations deployed for the agreed named systems; OMP re-run and --validate against the Customer Elastic project post-ingest; converted dashboards uploaded to Customer Kibana; assisted visualization polish for up to 15 OMP-Assisted panels (filter remaps, field-name fixes, query shape corrections); alert rule evaluation for up to 3 shortlisted rules (OMP-emitted rule adaptation, Elastic-native equivalent, or structured exit with written guidance)
• Phase 4: Side-by-side walkthrough comparing agreed Grafana dashboards to Kibana counterparts; enablement (ingest extension, converted dashboard operation); written residual backlog delivered
• Phase 5: Full artifact transfer; project close; close-out email

Countable scope boundaries (all hard caps):

• Ingest sample configurations: up to 6 EDOT/Elastic Agent configs and/or Prometheus remote-write endpoint(s)
• Assisted panel polish: up to 15 OMP-Assisted panels (internal delivery ceiling: 16 hours)
• Alert rule evaluation: up to 3 alert rules (internal delivery ceiling: 8 hours)

Out of scope

This engagement does not include:

• Full estate Agent rollout beyond the 6 sample configurations (requires change order)
• Historical Prometheus / Loki / Tempo data copy
• Panels beyond the 15-panel assisted allowance — these are documented in the written residual backlog
• Alert rules beyond the 3-rule evaluation allowance — these are documented in the written residual backlog
• Path 2 OTel remapping of Path 1 boards (separate engagement)
• Structural rebuild of panels OMP classifies as Outside (Outside panels do not consume the 15-panel allowance; they are documented in the residual backlog)
• Non-Prometheus datasources (Loki, CloudWatch, InfluxDB, etc.) unless explicitly listed in the Order Form
• Pixel-perfect Grafana visual parity
• Grafana-to-Kibana notification channel redesign (minimal notification path for alert validation only)
• Cutover from Grafana to Kibana — Customer remains dual-running Grafana throughout the engagement; cutover is a post-engagement Customer decision
• Log ingestion, APM/tracing, or RUM (available as follow-on engagements)
• Multi-cluster or multi-environment deployments beyond the agreed Elastic project

Any work outside the scope above requires a mutually executed change order (issued via the Order Form). All scope items not addressed within the fixed-price boundaries are documented in the written residual backlog at walkthrough and do not represent delivery shortfalls.

Start-gate requirements

The following must be confirmed before the engagement commences. Work may not begin until all start-gate artifacts are received and verified:

1. Grafana Dashboard JSON export — of the agreed representative sample, confirmed as available for OMP pre-assessment
2. Grafana Alert JSON export — covering the alert estate agreed in scope
3. OMP output / Emplify scorecard — from a pre-engagement run, reviewed before contract execution; the Assisted vs. Outside split must be communicated to the Customer
4. Target Elastic project confirmed — ECH 9.5+ or qualifying Observability Serverless with Prometheus remote-write/native PromQL (SA signoff required)

If start-gate artifacts are not in hand at kickoff, the delay is documented in writing and counts against consulting days in accordance with fixed-price terms.

Fixed-price conditions and assumptions

• Customer has an existing Elastic project (ECH 9.5+ or qualifying Observability Serverless) available before kickoff; SA has confirmed Prometheus remote-write / native PromQL support
• All four start-gate artifacts are received and verified before engagement commencement
• The representative dashboard and alert sample is agreed and fixed before kickoff; the panel allowance (15 Assisted) and alert rule allowance (3) have been communicated to and accepted by the Customer before signing
• A single Customer point of contact with decision-making authority is confirmed before kickoff
• Customer metrics are currently collected via Prometheus scrape, remote-write, OTel Collector, or Elastic Agent integrations compatible with Path 1 (Prometheus-family datasources)
• Customer provides ingest network access — no known blockers to landing metrics in the target Elastic project; security, proxy, or certificate constraints are surfaced before kickoff
• Customer remains on dual-run (Grafana live alongside Kibana) throughout the engagement; no Grafana cutover is expected within this package
• Customer technical resources (infrastructure/SRE engineers) are available for discovery sessions in Week 1 and walkthrough in Week 3
• The week-3 walkthrough date is confirmed at kickoff
• Delays caused by Customer unavailability, missing access, unresolved network constraints, or late artifact delivery count against consulting days and may affect delivery within the 3-week window

Deliverables

Kickoff & Planning

• One (1) post-kickoff written confirmation, documenting agreed scope, ingest approach, panel/alert caps, and walkthrough date
• One (1) Path, One (1) ingest plan, mapping agreed named systems to EDOT/Agent configurations or remote-write, locked in writing following discovery

Ingest & Validation (bounded to the named systems agreed in the Path 1 ingest plan)

• Prometheus remote-write endpoint(s) for the agreed named systems, deployed and verified in the Customer Elastic project 
• Up to Six (6) EDOT/Elastic Agent sample configurations for the agreed named systems, deployed and verified in the Customer Elastic project
• One (1) OMP re-run output and --validate results, confirming ingest against the Customer Elastic project post-deployment

Dashboards & Visualization

• In-scope Grafana dashboards converted and present in Customer Kibana, limited to the dashboards mapped in the Path One (1) ingest plan (flag below — recommend an explicit count)
• Assisted visualization polish applied to up to Fifteen (15) OMP-Assisted panels, limited to: filter remaps, field-name fixes, and query shape corrections (panel count and polish type applied documented per panel)

Alerting

• Up to Three (3) alert rules evaluated, with documented outcome per rule: tuned/validated OMP-emitted rule, Elastic-native equivalent built, or structured exit with written guidance

Walkthrough & Enablement

• One (1) side-by-side walkthrough session, scheduled and delivered live to the Customer, comparing agreed Grafana dashboards to their Kibana counterparts (flag below — "delivered or made available")
• One (1) enablement notes and configuration handoff document, covering ingest extension and converted dashboard operation

Closeout

• One (1) written residual backlog, itemizing panels beyond the Fifteen (15)-panel allowance, all out-of-scope panels, alert rules beyond the Three (3)-rule cap, and any non-migratable items — each with a recommended approach and effort class
• One (1) follow-on recommendations document (flag below — overlaps with residual backlog)
• Full artifact transfer: exports, OMP output, configurations, dashboard references, alert outcomes, and the residual backlog
• One (1) close-out email, confirming project completion and artifact transfer

Completion

The engagement will be considered complete upon: all in-scope deliverables produced; side-by-side walkthrough delivered or made available to Customer; and written residual backlog delivered. Elastic may declare Completion upon artifact transfer regardless of the Customer's subsequent Grafana cutover decision. Customer non-attendance at the walkthrough does not prevent Completion where Elastic has made the walkthrough available.

Completion

The engagement will be considered complete upon delivery of the final engagement report and Knowledge Transfer Workshop.

General terms

Capitalized terms used but not defined in this package description have the meanings given to them in the Order Form, the applicable Service Description, or the Elastic Services Agreement.

No formal Customer signoff and no milestone or deliverable acceptance process applies to this fixed-price Services Package. Each Service is deemed accepted upon Completion.

"Completion" means, with respect to a Service, the point at which Elastic has performed the activities within the scope of this Services Package, including (where applicable) delivery of any written report, materials, or knowledge transfer session described in this package's scope, or the expiry of the time-boxed engagement period for this package, whichever occurs first. Completion does not require Customer signoff or acceptance of any deliverable.

Customer is solely responsible for the backup, maintenance, and security of its environment, systems, and data. Elastic is not responsible for any configuration changes, data loss, downtime, or other impact to Customer's environment arising from or in connection with the Service.

This engagement delivers a configured net-new, self-managed Elastic document search solution, including search use case analysis, data modeling and ingestion, advanced example query implementation, advanced RBAC configuration, infrastructure provisioning, and a final report, within the defined fixed-price scope. Elastic does not guarantee any specific performance, cost, operational, or business outcome beyond the deliverables defined above. Customer remains solely responsible for ongoing operations, configuration changes made outside the engagement scope, and all post-engagement activities related to its environment.

References in this package to validation, testing, acceptance tests or acceptance test harnesses, golden queries, launch readiness, or go-live describe activities performed, demonstrated, or delivered by Elastic and do not create any Customer acceptance, signoff, testing, or payment condition. Completion occurs as defined in this package regardless of any such activity, and no portion of the fees is withheld or deferred pending any test result, validation, or go-live event.