Hosted log services that bill per gigabyte are the nearest Loggly alternatives, and open source log platforms you run yourself suit teams with the time to own them. Confirm how each one ingests, retains and alerts on your logs before you move anything.
You'd think swapping one hosted log tool for another is a weekend job: change the endpoint your shippers send to, and you're done. The endpoint is the easy bit. What takes the time is everything you've built on top of the logs over the years, and the billing model you've learned to live with. That's why this page starts with a checklist instead of a ranking.
Nine things to confirm before you move your logs
- Your shippers (syslog, a log agent, an OpenTelemetry collector or a language library) are supported by the new tool, or there's a documented replacement for each.
- You know your daily log volume on a normal day and on an incident day, in gigabytes, from your current account's usage page.
- The new tool's billing unit (ingest per gigabyte, daily volume, retention per gigabyte or a platform fee plus usage) has been mapped onto that volume.
- Searchable retention covers the debugging window you need, and there's a cheaper archive for anything kept longer for audit.
- The query language is one your team can write under pressure, whether that's Lucene-style search, SQL, PromQL or a vendor's own syntax.
- Every alert and saved search you rely on has been listed, because they rarely export, and each one has an owner who'll rebuild it.
- Live tail works at your volume without lag, since that's the screen you'll open first during an incident.
- You've decided whether logs stay alone or sit next to metrics and traces, so you don't buy a suite you won't open.
- Someone has agreed to own a self-run platform, with time set aside for upgrades and storage, before you choose the open source route.
Hosted log services
Better Stack Logs
Query logs and traces with SQL or PromQL in a hosted service that sits next to uptime and incident tools from the same company.
Logz.io
Managed observability on OpenSearch, combining log management, tracing and AI help for making sense of alerts.
Sematext Logs
Elasticsearch-based hosted log analytics, sized by how much you send each day and how long you keep it.
Axiom
For machine data at scale, an event store with its own query language and schema-less ingest, so you needn't define fields up front.
Coralogix
Telemetry lands in your own storage bucket with this observability platform, which changes the maths on long retention.
Datadog Log Management
Ingest, indexing, live tail and archiving of logs as one part of the broader Datadog monitoring suite.
Open source you can run yourself
Graylog
Central log management with an open edition you host, plus enterprise and security editions for larger needs.
SigNoz
Logs, traces and metrics in one open source platform, which you can run on your own servers or use as a cloud service.
OpenObserve
Open source observability for logs, metrics and traces, which its makers say needs far less storage than an Elastic setup.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Volume, retention and the shape of the bill
Items two to four are where the money is. Log tools usually bill per gigabyte ingested, and often charge separately for how long the data stays searchable. Some price by daily volume, some add a platform fee before usage, and some, like a self-run open source platform, swap the meter for your own servers and staff time. Each vendor's pricing page carries its current figures, and because the units differ, the useful work is converting each model into one number for your own volume.
Take your incident-day volume, not the average, and price a month at that rate, because that's the month that surprises you. Then go service by service and ask who reads each stream. Load balancer access logs, health checks and verbose debug output are often shipped by default and searched by nobody; filtering them at the shipper costs nothing and shrinks every quote you'll get. Where long retention is the sticking point, look at where the data physically lives: Coralogix keeps telemetry in your own storage bucket, which turns retention into a storage question rather than a log-vendor one.
Shippers, query language and live tail
Shippers come first because they decide how much code you touch. If you already send syslog or use an OpenTelemetry collector, switching is often a change of endpoint and API key. If you rely on a vendor-specific library inside your app, budget time to replace it in every service. I'd use the move to standardise on OpenTelemetry where you can, since it keeps your next move cheap too.
The query language matters more than it looks. During an outage nobody wants to read syntax docs. Have the person who's on call write three real queries in each finalist: a request ID lookup, a count of errors by service over the last hour, and a search across two fields. Whichever tool they can do that in fastest, without help, should rank higher than one with a longer feature list. Do the same for live tail at your real volume: open it, generate traffic, and watch for lag.
Alerts, saved searches and the logs-plus-metrics choice
Item six is the slow part of any log migration, so treat it as a clear-out as well as a move. List every alert, saved search and dashboard with the date it last fired or was opened. Anything untouched for months gets deleted rather than rebuilt. What's left is usually a much shorter list, and each survivor gets a named owner who rebuilds it and signs it off in the new tool.
On item eight: if your team only reads logs, a hosted log service keeps things simple and cheaper to reason about. If you're already chasing slow requests across services, a platform that links traces, metrics and logs saves real time, and the guides to observability tools and APM tools cover that side in more depth.
Running it yourself
Item nine decides the open source route. A self-run log platform removes the per-gigabyte meter and keeps every line on your own infrastructure, which matters when data location is a rule rather than a preference. In exchange, you own the storage plan, the upgrades, the index or compaction settings and the uptime of the very tool you'd use to debug an outage. I'd only go that way with a named owner and an hour or two a week set aside for them. If that person doesn't exist, a hosted service is cheaper in practice, even when its bill looks higher on paper.
Do you make log management or observability platforms? Tool makers can ask to be considered for a featured place on this list.
Want your tool on this list? Featured spots and reviews here are paid partnerships. Here's how to partner with me.


