Skip to content
Logmetry
What you run
SplunkMicrosoft SentinelDatadogDynatraceSolarWindsCriblOpenTelemetryA Data Lake
The Logmetry Blueprint
The whole methodPhase 01 · FoundationPhase 02 · FootprintPhase 03 · AI triageEverything as codeAI native
CompareJournalAboutBook the review

What you run · Dynatrace

Deep tracing where it earns its place

Dynatrace is the deep-tracing platform: OneAgent installs once per host and instruments everything automatically, PurePath traces follow every request end to end, and its Davis engine turns the topology it discovers into problems rather than raw alerts. It is licensed per host, so the ordinary node running basic checks pays the same rate as the crown-jewel path that genuinely needs causal depth. We keep Dynatrace on the paths that earn it, and run everything else on an open fleet you own.

01Today

How a Dynatrace estate looks today

A Dynatrace estate runs OneAgent everywhere at one per-host rate, so ordinary nodes on basic checks pay the same price as the crown-jewel paths that need causal depth.

How a Dynatrace estate looks todayOneAgent runs on every host with deep instrumentation by default, and every host pays the same per-host rate whether it is a crown-jewel application that needs causal analysis or an ordinary node reporting CPU, disk, and process checks. Renewals grow with the estate either way.YOUR ESTATEA few crown-jewel appsOrdinary nodes, basic checksONEAGENT, ON EVERY HOSTDeep instrumentation by defaultThe same agent, the same ratePER HOSTDYNATRACELicensed per host, at one rateCAUSAL ANALYSISGenuinely best in class,where depth is neededDASHBOARDSTopology discoveredautomaticallyOrdinary nodes use none of the depth,and pay for all of itWHAT YOU PAY FORCrown jewels and ordinarynodes pay the same rate.Renewals grow with theestate, either way.Depth a few paths need,priced on all of them.
How a Dynatrace estate looks today. One agent, one rate: the ordinary nodes that only ever report basic checks pay the same per-host price as the paths that genuinely need causal depth.

What Dynatrace does best

Dynatrace does depth like nobody else: automatic instrumentation, topology discovery, and causal analysis that genuinely shortens investigations on complex distributed applications. On the services where a minute of latency is revenue, that depth earns its price.

The problem is that per-host pricing does not distinguish the crown jewels from the crowd. Ordinary nodes running ordinary processes, watched by basic checks, pay the same rate as the paths that need causal AI, and renewals grow with the estate whether or not those nodes produce anything anyone reads.

02The five verbs

What we do to a Dynatrace estate

Keep it and govern what flows in, shrink its footprint, cut what it costs, extend it with an agent, and replace only where replacement is honest.

The Logmetry Blueprint applied to a Dynatrace estateCrown-jewel applications keep OneAgent and causal analysis, scoped and defensible. Ordinary nodes move to the open collector with every alert rebuilt and proven in writing first, feeding a metric store you own with no per-host bill. Everything lands at full fidelity in a lake in your own storage, and an agent you own follows the trace ID Dynatrace raises into the history around it. Pins: replace on the ordinary nodes, cut on the metric store, shrink on the Dynatrace boundary, keep on causal analysis, extend on the agent.YOUR ESTATEA few crown-jewel appsOrdinary nodes, basic checksONEAGENT ON CROWN JEWELSThe open collector everywhere elseAlert parity proven in writing firstRAll of itYOUR GROUNDTHE CONTROL LAYERRouted by tier,enriched in flightAs code, in your reposYOUR METRIC STOREThe ordinary tier, offper-host pricing entirelyCEverything, full fidelityTHE LAKE, YOURSPER HOSTOnly the deep tier crossesDYNATRACEScoped to the paths that earn itSCAUSAL ANALYSISKept, smaller, defensibleKThe alertYOUR AGENTFollows the trace ID intothe history around itEIt queries your historyWHAT CHANGEDThe deep tier stays, smallerand defensible.Ordinary nodes come offper-host pricing entirely.Every alert rebuilt andproven, tier by tier.
The Logmetry Blueprint applied. Deep tracing stays where it earns its price, the ordinary tier collects on an open fleet into a store you own, and the per-host meter reads only the paths that need depth. The pins mark where each of the five verbs acts.
KKeep

Dynatrace on the crown-jewel paths: the complex, revenue-critical applications where its causal depth and PurePath tracing genuinely shorten incidents and earn the per-host price.

SShrink

The licensed footprint contracts to the hosts that earn it, deliberately scoped and defended in the renewal conversation with the tier split written down before anyone negotiates.

CCut

The ordinary tier comes off per-host pricing entirely, onto OpenTelemetry collection and an open source time series database you own, running in your own account.

EExtend

Traces and logs land in the Lake carrying the same trace ID, so an Agent you own follows what Dynatrace raises into everything around it, on the ordinary tier Dynatrace never saw.

RReplace

Only the per-node infrastructure monitoring tier, and only with every alert rebuilt and proven to behave as it does today, tier by tier and in writing, first.

The letters mark where each verb acts in the drawing above

03The approach

Scope the depth, own the rest

The estate splits into the few hosts that need causal tracing and the many that need basic checks, and only one of those tiers should be rented.

We inventory the estate against what each tier actually consumes: which applications use the deep instrumentation, which hosts only ever report CPU, disk, and process checks. The ordinary tier moves to the open fleet with alert parity in writing. The deep tier stays, smaller and defensible.

What you own after: collection, alerts, and dashboards on the ordinary tier as code in your repositories, a Lake that holds the history, and a Dynatrace contract scoped to the work only Dynatrace can do.

04Questions

Asked about Dynatrace estates

Why not replace Dynatrace entirely?

Because on the right services its causal analysis is genuinely hard to reproduce, and pretending otherwise would cost you real capability. The honest move is scoping it to the paths that earn its price and taking the ordinary tier off per-host pricing.

What proves the open fleet matches our current alerting?

Every monitor and alert on the hosts that move is exported, rebuilt on the open stack, and proven to fire on the same conditions, tier by tier and in writing, before licensing changes. Coverage is the phase gate, not a hope.

What does the ordinary tier run on afterwards?

The OpenTelemetry Collector under OpAMP fleet control, feeding an open source time series database you own, with alerts and dashboards as code in your repositories. No per-host or per-agent licence anywhere in that path, and no lock-in to whichever store we build on.

Start with the review

You share your diagrams, we review them with you, and you leave with your version of the Logmetry Blueprint drawn on your Dynatrace estate. No system access, no obligation.

Book the reviewOr take the Logmetry Blueprint as a PDF →