FAQ
Frequently asked questions
Serious questions about observability migration deserve direct answers.
Will we lose historical data?
No. ExitGraph maintains read-only access to your legacy platform during and after migration. Critical historical data is exported and archived in object storage with configurable retention. We recommend maintaining legacy read access for 6-12 months post-cutover.
Do we need to re-instrument our services?
In most cases, no. The OTel collector acts as a translation layer, receiving telemetry from existing agents (Datadog Agent, Splunk Universal Forwarder) without code changes. You can migrate to native OTel instrumentation at your own pace after the cutover.
How does rollback work?
Every phase is rollback-safe. During mirror mode, your legacy system continues to receive all telemetry. Rollback is a configuration change — you switch the OTel collector routing back to legacy-only. No data migration, no configuration drift.
What happens to alert parity?
Every monitor and alert is inventoried, mapped to the target platform, and validated before cutover. Critical alerts are tested with synthetic data and compared against legacy behavior. Your team signs off on parity before any reads are switched.
How much can actually be translated automatically?
Typically 70-85% of dashboards and alerts translate automatically. The remainder requires manual review — usually vendor-specific widgets, complex query patterns, or custom integrations. ExitGraph provides a parity report with confidence scores for every asset.
How long does a full cutover take?
8-16 weeks depending on the number of services and source complexity. The assessment phase takes 2-3 weeks and produces a detailed timeline. Mirror mode can run as long as needed — there is no pressure to switch reads before you are ready.
Can we keep archive data queryable?
Yes. ExitGraph configures tiered storage with hot, warm, and cold tiers. Archive data lives in S3-compatible object storage with configurable retention and query-on-demand access through the target platform.
What if we use both Datadog and Splunk?
ExitGraph supports parallel migrations. Many organizations use Datadog for APM/metrics and Splunk for log search. Each is migrated independently with its own workspace, timeline, and parity plan.
What does the AI copilot actually do?
The copilot is grounded in your workspace data — services, alerts, runbooks, incidents, parity plan, and cutover state. It answers questions about your migration, identifies risks, summarizes status for executives, and helps investigate incidents using your actual operational context. It is not a generic chatbot.
How is pricing structured?
The assessment is a fixed-fee engagement starting at $15,000. Migration execution pricing depends on scope and complexity. A detailed cost estimate is included in the assessment deliverables.
Can we do this ourselves?
You can. Many teams attempt self-managed migrations. ExitGraph accelerates the process with structured tooling, automated translation, parity validation, and rollback safety that would take months to build internally. The assessment alone typically saves 4-6 weeks of discovery work.