SigNoz vs Grafana

Single Application vs Composable Observability

Grafana started as a data visualization tool. It slowly evolved into a tool that can take data from multiple data sources for visualization.

For observability, Grafana offers the LGTM stack (Loki for logs, Grafana for visualization, Tempo for traces, and Mimir for metrics). You need to configure and maintain multiple configurations for a full-stack observability setup.

The spread of tooling and configs due to different backends for each signal impacts developer experience and has a steeper learning curve.

SigNoz is powered by a single columnar datastore, ClickHouse. And SigNoz is OpenTelemetry-native. It’s easier to set up collection of all types telemetry data supported by OpenTelemetry and send it to SigNoz for visualization.

Ingestion, storage, querying, and visualization are optimized for ease of use and intelligent out-of-box correlation between the three signals.

For open-source and self-hosted users of SigNoz, there is less operational overhead and a better developer experience because of using a single data store.

SigNoz optimizes OTel attributes for faster querying, Loki was not built to index and query high-cardinality data

SigNoz allows you to index any kind of OpenTelemetry attribute. By default, we index all resource attribute regardless of its data type and cardinality.

Loki only indexes some of the low cardinality attributes, and the default limit of these resource attributes is 15. (Source)

In Loki, while converting attribute values in OTLP to index label values or structured metadata, any non-string values are converted to a string. While SigNoz supports indexing of attributes in string, number, and boolean. For example, in SigNoz, we can index duration, which makes querying and aggregating data faster. In Loki, if you want to perform an aggregation on such data, it happens over non-indexed data.(Source)

Loki, by design, is optimized for cost-effective log aggregation and storage rather than high-performance indexing. It doesn't support full-text indexing or advanced indexing on high-cardinality data like some other systems (e.g., Elasticsearch).

Instead, Loki focuses on indexing only labels (tags or metadata), making it ideal for scenarios where structured queries based on metadata are sufficient. So, for normal cases, it will use the labels, which are streams, as the main filter and then filter on the log data that is stored.

"As a Loki user or operator, your goal should be to use the fewest labels possible to store your logs." (Source: Grafana)

We did a logs performance benchmark of open-source SigNoz with Elasticsearch and Loki. Our key findings for Loki showed:

• Ingestion is mainly limited by the number of streams that it can handle.
• Loki was not able to return the results of our test queries.
• Loki took the least amount of storage as it indexed only two keys.

SigNoz is able to perform fast aggregation queries and also has efficient resource utilization during ingestion.

SigNoz offers a better dev experience for creating complex aggregations

SigNoz allows you to query any attribute and create complex aggregations on it. Our datastore, ClickHouse is built to support aggregations over massive datasets.

SigNoz allows querying and aggregation with a simple query builder where you can create any analytical query with just a few clicks.

Here's a quick demo of filtering for traces coming from a particular environment, then grouping them by k8s pod name to calculate latencies of spans from these pods.

In Grafana, you have to learn different query languages for different signals. For example, LogQL for logs, and traceQL for traces.

SigNoz uses columnar database for faster ingestion & aggregation

SigNoz uses ClickHouse - a fast open-source distributed columnar database for all three types of signals - logs, metrics, and traces. It was built to do analytical queries like `Group By` fast. Read more on what makes ClickHouse so fast →

Ingestion and aggregations are lightning-fast while providing best-in-class compression for economical storage.

SigNoz is much easier to self-host

If you want a self-hosted solution, SigNoz is a better choice. Since Grafana has multiple backends for different telemetry signals, it's difficult to manage. With SigNoz, you only need to manage a single backend for a full-stack observability setup. We also provide managed self-host services.