Important: This documentation is about an older version. It's relevant only to the release noted, many of the features and functions have been updated or replaced. Please view the current version.
Apache Parquet block format
Tempo has a default columnar block format based on Apache Parquet. This format is required for tags-based search as well as TraceQL, the query language for traces. The columnar block format improves search performance and enables a large ecosystem of tools to access the underlying trace data.
For more information, refer to the Parquet design document and Issue 1480. Additionally, there is now a Parquet v3 design document.
If you install using the new Helm charts, then Parquet is enabled by default.
Considerations
The Parquet block format is enabled by default in Tempo 2.0. No data conversion or upgrade process is necessary. As soon as the format is enabled, Tempo starts writing data in that format, leaving existing data as-is.
Block formats based on Parquet require more CPU and memory resources than the previous v2
format but provide search and TraceQL functionality.
Choose a different block format
The default block format is vParquet2
which is the latest iteration of Tempo’s Parquet based columnar block format. It is still possible to use the previous format vParquet
. To enable it, set the block version option to vParquet
in the Storage section of the configuration file.
# block format version. options: v2, vParquet, vParquet2
[version: vParquet]
It is possible to disable Parquet and use the previous v2
block format. This disables all forms of search, but also reduces resource consumption, and may be desired for a high-throughput cluster that does not need these capabilities. Set the block version option to v2
in the Storage section of the configuration file.
# block format version. options: v2, vParquet, vParquet2
[version: v2]
To re-enable the default vParquet2
format, remove the block version option from the Storage section of the configuration file or set the option to vParquet2
.
Parquet configuration parameters
Some parameters in the Tempo configuration are specific to Parquet. For more information, refer to the storage configuration documentation.
Trace search parameters
These configuration options impact trace search.
Parameter | Default value | Description |
---|---|---|
[read_buffer_size_bytes: <int>] | 10485676 | Size of read buffers used when performing search on a vParquet block. This value times the read_buffer_count is the total amount of bytes used for buffering when performing search on a Parquet block. |
[read_buffer_count: <int>] | 32 | Number of read buffers used when performing search on a vParquet block. This value times the read_buffer_size_bytes is the total amount of bytes used for buffering when performing search on a Parquet block. |
The cache_control
section contains the follow parameters for Parquet metadata objects:
Parameter | Default value | Description |
---|---|---|
[footer: | false | Specifies if the footer should be cached |
[column_index: <bool> | default = false] | false | Specifies if the column index should be cached |
[offset_index: <bool> | default = false] | false | Specifies if the offset index should be cached |
Convert to Parquet
If you have used an earlier version of the Parquet format, you can use tempo-cli
to convert a Parquet file from its existing schema to the one used in Tempo 2.0.
For instructions, refer to the Parquet convert command documentation.