Observability pricing comparison: Datadog, New Relic, Grafana, Dynatrace, Elastic and self-hosted log costs
Price one logs, metrics and traces profile across six observability options at list price, see what each leaves out, and what a 30% cut in log volume saves on each.
How to verify this
Open your browser's network tab and change an input: nothing is sent. The page computes with the formulas and the dated price table below. Some entries in that table could not be checked against an official page and are labelled. The same engine answers the API and the MCP tool, so you can reproduce any scenario from code.
Lowest list price for your profile: Elastic Cloud at $865.49/month. Highest: Datadog at $13,270.
- Note:Totals are list prices; contracts at this volume are usually negotiated.
- Heads-up:Self-hosted (Loki-style stack) is not verified against an official vendor price page (modelled or third-party values). Treat it as an estimate and get a quote.
- Heads-up:Self-hosted (Loki-style stack): only part of this pricing is published or it is modelled, so the total has low confidence.
- Note:Self-hosted operations are assumed at 0.25 FTE ($3750 per month). Set your own figure; it is usually the largest self-hosted cost.
- Note:Vendors that bill hosts, spans or metrics by data volume or by datapoint are priced with planning assumptions (5 GB per host per month; 1 GB per million spans; 4.32 GB per 1,000 series per month; 43200 datapoints per series per month; 16 GiB per APM host). No vendor publishes these figures; replace them with measured volumes.
- Note:Datadog does not include: Indexed spans beyond the 1 million per APM host that the host fee includes; Flex Logs storage and compute, and any retention beyond 30 days; Log archives in your own object storage and rehydration (scanned GB is billed per GB); Infrastructure Enterprise, APM Pro and APM Enterprise tiers, on-demand and month-to-month rates; Per-seat products (On-Call, Incident Management) and other products such as RUM, Synthetics and Security.
- Note:New Relic does not include: Data Plus option ($0.60 per GB), the EU data centre surcharge, and Enterprise terms; Core and Basic users, and the Standard edition's first-user pricing; Compute (query) charges listed on the pricing page; Extended retention for traces and metrics, and the 100 GB free tier is applied once across all data types; Real infrastructure, trace and metric volumes: they are estimated from planning assumptions.
- Note:Grafana Cloud does not include: Volume discounts (the 'starts at' rate is used throughout); Series generated by infrastructure integrations beyond the custom metrics you enter; Incident Response and On-Call users ($20), Grafana Assistant ($20) and Enterprise plugins ($55); Log query volume above the fair-use ratio, and support plans; Non-Kubernetes hosts are priced at the Kubernetes Monitoring host rate because the page lists no separate host plan.
- Note:Dynatrace does not include: Log query charges ($0.0035 per GiB scanned) and the retain option with included queries; Metric and trace retention, trace queries, Kubernetes Platform Monitoring (per pod hour) and Application Security; Foundation & Discovery hosts, and subscription commitments or negotiated discounts; Real host memory: 16 GiB per APM host is assumed.
- Note:Elastic Cloud does not include: Support plans (Gold 5%, Platinum 10%, Enterprise 15% of consumption; Standard is included); Data transfer out beyond 50 GB free ($0.05 per GB), synthetics, Managed LLM, workflows and cross-project search; Retention of traces and metrics, and Elastic Cloud Hosted (resource-based) pricing; Real infrastructure, trace and metric volumes: they are estimated from planning assumptions.
- Note:Self-hosted (Loki-style stack) does not include: Metrics, traces, dashboards, alerting and on-call tooling; Object storage requests, data transfer, backups, caches (memcached), load balancers and multi-zone redundancy; Query headroom beyond Loki's p90 sizing, reserved-instance or savings-plan discounts, and the negotiated discount (never applied here); Security, upgrade and on-call effort beyond the FTE you enter.
Lowest monthly total
$865.49
Elastic Cloud per month, $10,386 a year
Highest monthly total
$13,270
Datadog per month
Logs ingested per month
6,000 GB
3,600M events indexed, 4,200 GB archived
Saved by 30% less logs
$23.29 to $2,016
per month, across the selected vendors
Monthly cost by component, per vendor
- Log ingest
- Log indexing and retention
- Infrastructure
- APM and traces
- Metrics
- Seats
- Other
| Vendor | Log ingest | Log indexing and retention | Infrastructure | APM and traces | Metrics | Seats | Other | Monthly total | Annual total | Effective per GB | 30% less logs saves | Confidence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Elastic Cloud | $540.00 | $256.50 | $23.00 | $45.00 | $0.99 | $0.00 | $0.00 | $865.49 | $10,386 | $0.1442 | $238.95 (27.6%) | Medium |
| Self-hosted (Loki-style stack) * (logs only) | $280.32 | $77.63 | $0.00 | $0.00 | $0.00 | $0.00 | $3,750 | $4,108 | $49,295 | $0.6847 | $23.29 (0.6%) | Low |
| Grafana Cloud | $2,678 | $840.00 | $2,891 | $1,103 | $0.00 | $136.00 | $19.00 | $7,666 | $91,996 | $1.28 | $1,062 (13.9%) | Medium |
| New Relic | $2,360 | $420.00 | $400.00 | $200.00 | $17.28 | $6,980 | $0.00 | $10,377 | $124,527 | $1.73 | $846.00 (8.2%) | Medium |
| Dynatrace | $1,118 | $264.03 | $4,380 | $5,933 | $648.00 | $0.00 | $0.00 | $12,343 | $148,113 | $2.06 | $414.49 (3.4%) | Medium |
| Datadog | $600.00 | $6,120 | $5,000 | $1,550 | $0.00 | $0.00 | $0.00 | $13,270 | $159,240 | $2.21 | $2,016 (15.2%) | High |
Self-hosted: unverified *Self-hosted: low confidence
| Vendor | Price basis | Not included in the total |
|---|---|---|
| Elastic Cloud | Elastic Cloud Serverless, Observability Complete tier. The page shows 'as low as' prices, so the rate you get may be higher. | Support plans (Gold 5%, Platinum 10%, Enterprise 15% of consumption; Standard is included); Data transfer out beyond 50 GB free ($0.05 per GB), synthetics, Managed LLM, workflows and cross-project search; Retention of traces and metrics, and Elastic Cloud Hosted (resource-based) pricing; Real infrastructure, trace and metric volumes: they are estimated from planning assumptions |
| Self-hosted (Loki-style stack) | Modelled, not a vendor price list. Logs only. AWS on-demand rates in us-east-1; node capacity derived from Loki's published sizing; compression is an assumption. | Metrics, traces, dashboards, alerting and on-call tooling; Object storage requests, data transfer, backups, caches (memcached), load balancers and multi-zone redundancy; Query headroom beyond Loki's p90 sizing, reserved-instance or savings-plan discounts, and the negotiated discount (never applied here); Security, upgrade and on-call effort beyond the FTE you enter |
| Grafana Cloud | Pro plan 'starts at' prices. The page states volume discounts apply automatically at scale but does not publish the tiers, so large-volume totals here are overstated. | Volume discounts (the 'starts at' rate is used throughout); Series generated by infrastructure integrations beyond the custom metrics you enter; Incident Response and On-Call users ($20), Grafana Assistant ($20) and Enterprise plugins ($55); Log query volume above the fair-use ratio, and support plans; Non-Kubernetes hosts are priced at the Kubernetes Monitoring host rate because the page lists no separate host plan |
| New Relic | Original data option at list price. Full platform user price is the annual-commitment price. Billing is per GB ingested for every data type. | Data Plus option ($0.60 per GB), the EU data centre surcharge, and Enterprise terms; Core and Basic users, and the Standard edition's first-user pricing; Compute (query) charges listed on the pricing page; Extended retention for traces and metrics, and the 100 GB free tier is applied once across all data types; Real infrastructure, trace and metric volumes: they are estimated from planning assumptions |
| Dynatrace | Dynatrace Platform Subscription rate card, list price in USD. The page says discounts are available on request. | Log query charges ($0.0035 per GiB scanned) and the retain option with included queries; Metric and trace retention, trace queries, Kubernetes Platform Monitoring (per pod hour) and Application Security; Foundation & Discovery hosts, and subscription commitments or negotiated discounts; Real host memory: 16 GiB per APM host is assumed |
| Datadog | Annual-commitment list prices from the Billed Annually column. On-demand and month-to-month columns are higher (for example Infrastructure Pro $18 per host). | Indexed spans beyond the 1 million per APM host that the host fee includes; Flex Logs storage and compute, and any retention beyond 30 days; Log archives in your own object storage and rehydration (scanned GB is billed per GB); Infrastructure Enterprise, APM Pro and APM Enterprise tiers, on-demand and month-to-month rates; Per-seat products (On-Call, Incident Management) and other products such as RUM, Synthetics and Security |
| Assumption | Value | Unit |
|---|---|---|
| gb per host month | 5 | GB per host per month |
| gb per million spans | 1 | GB per million spans |
| gb per 1000 series month | 4.32 | GB per 1,000 series per month |
| datapoints per series month | 43,200 | datapoints per series per month |
| host memory gib | 16 | GiB per APM host |
* Not verified against an official vendor page: the figures are modelled or come from a third party. Totals are list prices.
Monthly cost per vendor as the sweep input changes
Calculating the range…
How it is calculated
Monthly volumes
G_ingest = 30 × L × (1 − d) E_indexed = G_ingest × i × 10^9 ÷ (bytes per event × 10^6) (million events) G_archive = G_ingest × (1 − i) L = log GB per day, d = dropped share, i = indexed share, 1 GB = 10^9 bytes
Vendor cost
C_vendor = (1 − δ) × Σ over components of tiered(max(q_c − free_c, 0), rate_c) δ = negotiated discount. q_c is the quantity in the component's unit: GB ingested, million events indexed (rate chosen by retention tier), hosts, containers, custom metrics, APM hosts, spans, seats or GB stored. Shared allowances (New Relic's 100 GB) are used up once, in component order.
Stored volume and retention
S = L × (1 − d) × [ i × indexed days + (1 − i) × archive days ] (GB held in steady state) Vendors that charge per extra retention period (New Relic, Grafana Cloud) bill GB ingested × ceil((days − included days) ÷ period length), for the indexed and archived shares.
Self-hosted
storage = S × compression ratio × storage $/GB-month compute = ceil(L × (1 − d) ÷ node GB-per-day capacity) × hourly $ × 730 operations = FTE × loaded engineer $ per year ÷ 12
Effective rate and what-if
effective $/GB = monthly total ÷ G_ingest what-if = every vendor recomputed with G_ingest × 0.7; saving = total − what-if total
Assumptions
- Everything is list price in USD. Datadog uses the annual-commitment column; Grafana Cloud 'starts at' prices and Elastic 'as low as' prices are used as published. Taxes, support plans and cloud egress are not included.
- Each total covers the line items the vendor publishes. The 'What each price leaves out' table beside the results lists what is missing for every vendor, and the same text appears as a warning.
- Vendors that bill hosts, spans or metric series by data volume or datapoints (New Relic, Elastic, Grafana traces, Dynatrace) are priced with planning assumptions: 5 GB per host per month, 1 GB per million spans, 4.32 GB per 1,000 series and 43,200 datapoints per series per month, 16 GiB per Dynatrace APM host. No vendor publishes these; they are marked unverified and shown with the result.
- Only vendors that price indexing separately (Datadog) charge less for logs you archive instead of index. The others bill every ingested GB, and storage-based vendors also bill the archive period.
- The indexed share is kept for the indexed retention, the rest for the archive retention. A vendor that lists no price for your retention is priced at its longest listed tier and flagged.
- Dynatrace is billed in GiB (1 GiB = 1.0737 GB) and by the hour (730 hours a month). A month is 30 days.
- Self-hosted sizes a Loki-style log pipeline only: object storage, enough 8-vCPU nodes for the daily volume, and operations time. Its AWS prices are third-party or undated, so it is marked unverified. The negotiated discount is never applied to it.
- The ranking is by monthly total. The cheapest row is cheapest for the line items it includes, which is not the same for every vendor, so read the leaves-out table before you decide.
The data behind it
| Vendor | Price basis | Confidence | Checked | Gotchas |
|---|---|---|---|---|
| Datadog | Annual-commitment list prices from the Billed Annually column. On-demand and month-to-month columns are higher (for example Infrastructure Pro $18 per host). | High | 2026-11-05 | Indexing is priced per million events and usually dominates the log bill; the rate rises with retention. The list page prices indexing for 3 to 30 days only; longer retention needs Flex Logs, which is not modelled. Custom metric and container allotments are pooled per account (Pro: 100 metrics and 5 containers per host). Annual-commitment prices are used; the on-demand rate is higher. |
| New Relic | Original data option at list price. Full platform user price is the annual-commitment price. Billing is per GB ingested for every data type. | Medium | 2026-11-05 | Every GB ingested is billed whether or not anyone searches it; there is no cheaper archive tier in this model. Infrastructure, trace and metric data are billed per GB, so their cost depends on agent volume rather than host counts. Full platform users are priced per person; Core users ($49) and Basic users (free) are priced differently and not modelled. Retention beyond the default costs $0.05 per GB ingested for each extra 30 days. Data Plus costs $0.60 per GB instead of $0.40. |
| Grafana Cloud | Pro plan 'starts at' prices. The page states volume discounts apply automatically at scale but does not publish the tiers, so large-volume totals here are overstated. | Medium | 2026-11-05 | Listed rates are 'starts at' prices; the effective rate falls with volume but the tiers are not published. Logs have separate processed, written and retained rates; retention beyond 30 days is billed per extra 30-day increment. Users beyond the 3 included cost $8 each, and incident-response users are priced separately. Query volume above a fair-use ratio of 100 times written log volume is billed. |
| Dynatrace | Dynatrace Platform Subscription rate card, list price in USD. The page says discounts are available on request. | Medium | 2026-11-05 | Log queries are billed per GiB scanned on top of ingest and retain, so search habits change the bill. Full-Stack Monitoring bills memory-GiB-hours, so large APM hosts cost more than small ones. Metrics are billed per datapoint, so reporting frequency matters as much as series count. There are no per-seat fees. |
| Elastic Cloud | Elastic Cloud Serverless, Observability Complete tier. The page shows 'as low as' prices, so the rate you get may be higher. | Medium | 2026-11-05 | Prices are shown 'as low as'; your rate can be higher. Retention is billed on everything stored, so long archive periods add a monthly storage charge on all data. Metrics in time-series mode are billed at 25% of the standard ingest and retention rates. There are no per-host or per-seat fees, so cost follows data volume only. |
| Self-hosted (Loki-style stack) * | Modelled, not a vendor price list. Logs only. AWS on-demand rates in us-east-1; node capacity derived from Loki's published sizing; compression is an assumption. | Low | Not verified | Cost moves from licences into engineering time: at 0.25 FTE the people cost often exceeds the servers. Only the log pipeline is sized; metrics, traces, dashboards and alerting are not. Node capacity is a rough planning figure; heavy querying can need several times more. |
| Vendor | Item | Price | Included | Verified |
|---|---|---|---|---|
| Datadog | Log ingestion | $0.10 per GB ingested per month | none | Yes |
| Datadog | Log indexing | $1.06 (3 days), $1.27 (7 days), $1.70 (15 days), $2.50 (30 days) per million events indexed per month | none | Yes |
| Datadog | Infrastructure hosts (Pro) | $15.00 per host per month | none | Yes |
| Datadog | Containers beyond the allotment | $1.00 per container per month | 5 per host | Yes |
| Datadog | Custom metrics beyond the allotment | $0.05 per custom metric per month | 100 per host | Yes |
| Datadog | APM hosts | $31.00 per APM host per month | none | Yes |
| Datadog | APM span ingestion beyond the allotment | $0.10 per GB of traces per month | 150 per APM host | Yes |
| New Relic | Data ingest: logs | $0.40 per GB ingested per month | shares 100 free | Yes |
| New Relic | Extended log retention | $0.05 per GB per extra retention period per month | none | Yes |
| New Relic | Data ingest: infrastructure agents | $0.40 per GB of agent data per month | shares 100 free | Yes |
| New Relic | Data ingest: traces | $0.40 per GB of traces per month | shares 100 free | Yes |
| New Relic | Data ingest: custom metrics | $0.40 per GB of metric data per month | shares 100 free | Yes |
| New Relic | Full platform users (Pro) | $349.00 per seat per month | none | Yes |
| Grafana Cloud | Pro platform fee | $19.00 per account per month | none | Yes |
| Grafana Cloud | Logs: processed plus written | $0.45 per GB ingested per month | 50 free | Yes |
| Grafana Cloud | Logs: retention beyond 30 days | $0.10 per GB per extra retention period per month | none | Yes |
| Grafana Cloud | Metrics beyond 10k active series | $0.0065 per custom metric per month | 10,000 free | Yes |
| Grafana Cloud | Traces: processed plus written | $0.45 per GB of traces per month | 50 free | Yes |
| Grafana Cloud | Kubernetes Monitoring hosts | $7.20 per host per month | 3.1 free | Yes |
| Grafana Cloud | Kubernetes Monitoring containers | $0.50 per container per month | 52.7 free | Yes |
| Grafana Cloud | Application Observability hosts | $18.00 per APM host per month | none | Yes |
| Grafana Cloud | Grafana Visualization users beyond 3 | $8.00 per seat per month | 3 free | Yes |
| Dynatrace | Logs: ingest and process | $0.20 per GB ingested per month | none | Yes |
| Dynatrace | Logs: retain | $0.0007 per GB stored per day | none | Yes |
| Dynatrace | Infrastructure Monitoring hosts | $0.04 per host without APM per hour | none | Yes |
| Dynatrace | Full-Stack Monitoring (APM hosts) | $0.01 per GiB of APM host memory per hour | none | Yes |
| Dynatrace | Metrics: ingest and process | $0.15 per 100,000 datapoints per month | none | Yes |
| Dynatrace | Traces: ingest and process | $0.20 per GB of traces per month | none | Yes |
| Elastic Cloud | Ingest: logs | $0.09 per GB ingested per month | none | Yes |
| Elastic Cloud | Retention: logs | $0.019 per GB stored per month | none | Yes |
| Elastic Cloud | Ingest: infrastructure metrics | $0.023 per GB of agent data per month | none | Yes |
| Elastic Cloud | Ingest: traces | $0.09 per GB of traces per month | none | Yes |
| Elastic Cloud | Ingest: custom metrics | $0.023 per GB of metric data per month | none | Yes |
| Part | Value | Basis |
|---|---|---|
| Object storage, USD per GB-month | 0.02 | S3 standard storage in us-east-1, first 50 TB, from an AWS Solutions cost example that carries no date. The live AWS pricing page is rendered by script and could not be read, so this is not marked verified. |
| Stored size as a fraction of raw size | 0.25 | Stored size as a fraction of raw size (0.25 means 4 to 1). A third-party article puts gzip at 70 to 80 percent and snappy at 50 to 60 percent reduction; this is an assumption, not a measurement. |
| Node (m6i.2xlarge), USD per hour | 0.38 | m6i.2xlarge (8 vCPU, 32 GiB) on-demand $0.384 per hour, read from a third-party price index. AWS's own page could not be read. |
| Node capacity, GB of logs per day | 630 | Derived: Loki's sizing guide lists about 38 CPU cores (20 write, 16 read, 2 compactor) for under 3 TB per day, so one 8-vCPU node covers about 3,000 x 8 / 38 = 630 GB per day. Loki says these are rough p90 requests and queries can need much more. |
| Assumption | Value | Unit | Basis |
|---|---|---|---|
| gb per host month | 5 | GB per host per month | Planning assumption for agent data (system metrics, process data) per host. New Relic says ingest is measured per agent and publishes no figure; Elastic publishes none. Measure your own. |
| gb per million spans | 1 | GB per million spans | Planning assumption of 1,000 bytes per span. No vendor publishes an average span size. Measure your own. |
| gb per 1000 series month | 4.32 | GB per 1,000 series per month | Planning assumption: one sample per minute (43,200 per 30 days) at 100 bytes per sample. Measure your own. |
| datapoints per series month | 43,200 | datapoints per series per month | Planning assumption: one datapoint per series per minute for 30 days. |
| host memory gib | 16 | GiB per APM host | Planning assumption for Dynatrace Full-Stack Monitoring, which bills per memory-GiB-hour. Use your real average host memory. |
| Constant | Value |
|---|---|
| Hours per month (hourly rates) | 730 |
| Days per month (daily rates) | 30 |
| What-if log reduction (%) | 30 |
| Topic | What catches people out |
|---|---|
| Ingest vs indexing | Some vendors price log ingest and log indexing separately, and indexing usually dominates the bill. |
| Retention | Retention multiplies indexed cost: a longer retention tier has a higher rate, or an extra charge for every GB per additional period. |
| Cardinality | A single high-cardinality tag can multiply custom metric series by thousands, and series are what you are billed on. |
| Tracing | Trace sampling changes span volume, and span volume is what trace-priced vendors bill. |
| Rehydration | Rehydrating archived logs can bill again, once for scanning and again for indexing. |
| Seats | Seat-based vendors get costly with many responders; per-GB vendors may charge nothing for the same people. |
| Self-hosting | Self-hosting moves cost out of the vendor invoice and into engineering time, which is easy to under-count. |
| Discounts | Real contracts are negotiated; list price is the starting point, not the number you will pay. |
Last updated 2026-11-05 (dataset 2026-11, refreshed monthly, manually verified). Entries marked unverified or modelled could not be checked against an official page.
- https://www.datadoghq.com/pricing/list/
- https://www.datadoghq.com/pricing/
- https://docs.datadoghq.com/account_management/billing/custom_metrics/
- https://newrelic.com/pricing
- https://docs.newrelic.com/docs/data-apis/manage-data/manage-data-retention/
- https://grafana.com/pricing/
- https://grafana.com/docs/grafana-cloud/cost-management-and-billing/understand-your-invoice/logs-invoice/
- https://www.dynatrace.com/pricing/rate-card/
- https://www.dynatrace.com/pricing/
- https://www.elastic.co/pricing/serverless-observability
- https://www.elastic.co/docs/deploy-manage/cloud-organization/billing/elastic-observability-billing-dimensions
- https://grafana.com/docs/loki/latest/setup/size/
- https://docs.aws.amazon.com/solutions/latest/live-streaming-on-aws-with-amazon-s3/cost-example-1.html
- https://www.doit.com/compute/spot/us-east-1/m6i.2xlarge
- https://oneuptime.com/blog/post/2026-01-21-loki-storage-costs/view
Three ways to reduce this number
- Drop or sample noisy debug logs before ingest. It is the largest lever because every vendor bills what it receives.
- Index only what you search and archive the rest. Use the indexed-share sweep to see how much each point of indexing costs.
- Cap metric tag cardinality and set trace sampling rates, so custom metrics and spans stop growing with traffic.
Use it from code
The same engine is available as a REST endpoint and as an MCP tool named calc_observability_log_cost_comparator. Your scenario as an API call:
curl -X POST 'https://api.x12i.com/v1/calc/observability-log-cost-comparator' \
-H 'Authorization: Bearer YOUR_API_KEY' \
-H 'Content-Type: application/json' \
-d '{
"input": {
"log_gb_per_day": 200,
"avg_log_event_bytes": 500,
"indexed_share_pct": 30,
"drop_or_sample_pct": 0,
"retention_days_indexed": 15,
"archive_retention_days": 90,
"hosts": 200,
"containers_per_host": 15,
"custom_metrics": 10000,
"apm_spans_millions_per_month": 500,
"user_seats": 20,
"vendors": [
"datadog",
"new_relic",
"grafana_cloud",
"dynatrace",
"elastic_cloud",
"self_hosted"
],
"negotiated_discount_pct": 0,
"self_hosted_ops_fte": 0.25,
"loaded_engineer_usd_per_year": 180000,
"current_observability_pain": "none"
},
"options": {
"explain": false
}
}' Schema: GET /v1/calc/observability-log-cost-comparator/schema
Frequently asked questions
How much does Datadog log management cost per GB?
Datadog bills logs twice. Ingestion is $0.10 per GB. Indexing is priced per million events by retention; on the annual-commitment list price read on 5 November 2026 it was $1.06 for 3 days, $1.27 for 7, $1.70 for 15 and $2.50 for 30 days per million events (on-demand rates are higher). At 500 bytes per event a GB is about 2 million events, so indexing a GB for 15 days costs about $3.40 on top of the $0.10 ingest. Logs you only archive pay the ingest rate alone. The data table below shows the values the calculator uses.
Is self-hosting observability cheaper?
For logs it can be, but people cost is the swing factor. At 0.25 of an engineer and $180,000 loaded cost, operations alone are $3,750 a month before any server. This calculator sizes a Loki-style log pipeline only (storage, nodes, operations), so it does not include metrics, traces, dashboards or alerting. Self-hosting tends to win when log volume is high enough that per-GB vendor rates exceed that fixed cost, and when a platform team already exists. Change the effort and volume and use the sweep chart to find where the lines cross for your profile.
What is OpenTelemetry and does it reduce lock-in?
OpenTelemetry is an open, vendor-neutral standard with SDKs and a collector for producing logs, metrics and traces. If you instrument with it, the same telemetry can go to any compatible backend, so changing vendor or sending to two at once is a configuration change rather than a re-instrumentation. That reduces lock-in on the collection side; dashboards, alerts and query languages remain vendor-specific. The collector can also drop or sample data before it reaches a vendor, which is the biggest cost lever.