Structured logs should be shipped in a valid format. If the structure contains invalid characters for parsing, these should be stripped at the Agent level using the mask_sequences feature.
As a best practice, it is recommended to use at most 20 processors per pipeline.
Create custom grok rules to parse the full message or a specific attribute of your raw event. For more information, see the parsing section. As a best practice, it is recommended to use at most 10 parsing rules within a grok processor.
Click Parse my logs to kickstart a set of three parsing rules for the logs flowing through the underlying pipeline. Refine attribute naming from there, and add new rules for other type of logs if needed. This feature requires that the corresponding logs are being indexed, and actually flowing in—you can temporarily deactivate or sample down exclusion filters to make this work for you.
Select a sample by clicking on it to trigger its evaluation against the parsing rule and display the result at the bottom of the screen.
Up to five samples can be saved with the processor, and each sample can be up to 5000 characters in length. All samples show a status (match or no match), which highlights if one of the parsing rules of the grok parser matches the sample.
As Datadog receives logs, it timestamps them using the value(s) from any of these default attributes:
timestamp
date
_timestamp
Timestamp
eventTime
published_date
If your logs have dates in an attribute that are not in this list, use the log date remapper processor to define their date attribute as the official log timestamp:
If your logs don’t have a timestamp that conforms to the formats listed above, use the grok processor to extract the epoch time from the timestamp to a new attribute. The date remapper uses the newly defined attribute.
To see how a custom date and time format can be parsed in Datadog, see Parsing dates.
Notes:
Log events can be submitted up to 18 hours in the past and two hours in the future.
As of ISO 8601-1:2019, the basic format is T[hh][mm][ss] and the extended format is T[hh]:[mm]:[ss]. Earlier versions omitted the T (representing time) in both formats.
If your logs don’t contain any of the default attributes and you haven’t defined your own date attribute, Datadog timestamps the logs with the date it received them.
If multiple log date remapper processors are applied to a given log within the pipeline, the last one (according to the pipeline’s order) is taken into account.
Use the status remapper processor to assign attributes as an official status to your logs. For example, add a log severity level to your logs with the status remapper.
Strings beginning with emerg or f (case-insensitive) map to emerg (0)
Strings beginning with a (case-insensitive) map to alert (1)
Strings beginning with c (case-insensitive) map to critical (2)
Strings beginning with err (case-insensitive) map to error (3)
Strings beginning with w (case-insensitive) map to warning (4)
Strings beginning with n (case-insensitive) map to notice (5)
Strings beginning with i (case-insensitive) map to info (6)
Strings beginning with d, t, v, trace, or verbose (case-insensitive) map to debug (7)
Strings beginning with o or s, or matching OK or Success (case-insensitive) map to OK
All others map to info (6)
Note: If multiple log status remapper processors are applied to a log within a pipeline, only the first one in the pipeline’s order is considered. Additionally, for all pipelines that match the log, only the first status remapper encountered (from all applicable pipelines) is applied.
The service remapper processor assigns one or more attributes to your logs as the official service.
Note: If multiple service remapper processors are applied to a given log within the pipeline, only the first one (according to the pipeline’s order) is taken into account.
message is a key attribute in Datadog. Its value is displayed in the Content column of the Log Explorer to provide context on the log. You can use the search bar to find a log by the log message.
Use the log message remapper processor to define one or more attributes as the official log message. Define more than one attribute for cases where the attributes might not exist and an alternative is available. For example, if the defined message attributes are attribute1, attribute2, and attribute3, and attribute1 does not exist, then attribute2 is used. Similarly, if attribute2 does not exist, then attribute3 is used.
To define message attributes, first use the string builder processor to create a new string attribute for each of the attributes you want to use. Then, use the log message remapper to remap the string attributes as the message.
Note: If multiple log message remapper processors are applied to a given log within the pipeline, only the first one (according to the pipeline order) is taken into account.
The remapper processor remaps any source attribute(s) or tags to another target attribute or tag. For example, remap user by firstname to target your logs in the Log Explorer:
Constraints on the tag/attribute name are explained in the attributes and tags documentation. Some additional constraints, applied as : or ,, are not allowed in the target tag/attribute name.
If the target of the remapper is an attribute, the remapper can also try to cast the value to a new type (String, Integer or Double). If the cast is not possible, the original type is kept.
Note: The decimal separator for Double need to be ..
{"type":"attribute-remapper","name":"Remap <SOURCE_ATTRIBUTE> to <TARGET_ATTRIBUTE>","is_enabled":true,"source_type":"attribute","sources":["<SOURCE_ATTRIBUTE>"],"target":"<TARGET_ATTRIBUTE>","target_type":"tag","target_format":"integer","preserve_source":false,"override_on_conflict":false}
Parameter
Type
Required
Description
type
String
Yes
Type of the processor.
name
String
No
Name of the processor.
is_enabled
Boolean
No
If the processors is enabled or not. Default: false.
source_type
String
No
Defines if the sources are from log attribute or tag. Default: attribute.
sources
Array of strings
Yes
Array of source attributes or tags
target
String
Yes
Final attribute or tag name to remap the sources to.
target_type
String
No
Defines if the target is a log attribute or a tag. Default: attribute.
target_format
String
No
Defines if the attribute value should be cast to another type. Possible values: auto, string, or integer. Default: auto. When set to auto, no cast is applied.
preserve_source
Boolean
No
Remove or preserve the remapped source element. Default: false.
override_on_conflict
Boolean
No
Override or not the target element if already set. Default: false.
The user-agent parser processor takes a useragent attribute and extracts OS, browser, device, and other user data. When set up, the following attributes are produced:
Note: If your logs contain encoded user-agents (for example, IIS logs), configure this Processor to decode the URL before parsing it.
{"type":"user-agent-parser","name":"Parses <SOURCE_ATTRIBUTE> to extract all its User-Agent information","is_enabled":true,"sources":["http.useragent"],"target":"http.useragent_details","is_encoded":false}
Parameter
Type
Required
Description
type
String
Yes
Type of the processor.
name
String
No
Name of the processor.
is_enabled
Boolean
No
If the processors is enabled or not. Default: false.
sources
Array of strings
No
Array of source attributes. Default: http.useragent.
target
String
Yes
Name of the parent attribute that contains all the extracted details from the sources. Default: http.useragent_details.
is_encoded
Boolean
No
Define if the source attribute is url encoded or not. Default: false.
Use the category processor to add a new attribute (without spaces or special characters in the new attribute name) to a log matching a provided search query. Then, use categories to create groups for an analytical view (for example, URL groups, machine groups, environments, and response time buckets).
Notes:
The syntax of the query is the one in the Log Explorer search bar. This query can be done on any log attribute or tag, whether it is a facet or not. Wildcards can also be used inside your query.
Once the log has matched one of the processor queries, it stops. Make sure they are properly ordered in case a log could match several queries.
The names of the categories must be unique.
Once defined in the category processor, you can map categories to log status using the log status remapper.
Define the category processor on the Pipelines page. For example, to categorize your web access logs based on the status code range value ("OK" for a response code between 200 and 299, "Notice" for a response code between 300 and 399, ...) add this processor:
{"type":"category-processor","name":"Assign a custom value to the <TARGET_ATTRIBUTE> attribute","is_enabled":true,"categories":[{"filter":{"query":"<QUERY_1>"},"name":"<VALUE_TO_ASSIGN_1>"},{"filter":{"query":"<QUERY_2>"},"name":"<VALUE_TO_ASSIGN_2>"}],"target":"<TARGET_ATTRIBUTE>"}
Parameter
Type
Required
Description
type
String
Yes
Type of the processor.
name
String
No
Name of the processor.
is_enabled
Boolean
No
If the processors is enabled or not. Default: false
categories
Array of Object
Yes
Array of filters to match or not a log and their corresponding name to assign a custom value to the log.
target
String
Yes
Name of the target attribute which value is defined by the matching category.
Use the arithmetic processor to add a new attribute (without spaces or special characters in the new attribute name) to a log with the result of the provided formula. This remaps different time attributes with different units into a single attribute, or compute operations on attributes within the same log.
An arithmetic processor formula can use parentheses and basic arithmetic operators: -, +, *, /.
By default, a calculation is skipped if an attribute is missing. Select Replace missing attribute by 0 to automatically populate missing attribute values with 0 to ensure that the calculation is done.
Notes:
An attribute may be listed as missing if it is not found in the log attributes, or if it cannot be converted to a number.
When using the operator -, add spaces around it because attribute names like start-time may contain dashes. For example, the following formula must include spaces around the - operator: (end-time - start-time) / 1000.
If the target attribute already exists, it is overwritten by the result of the formula.
Results are rounded up to the 9th decimal. For example, if the result of the formula is 0.1234567891, the actual value stored for the attribute is 0.123456789.
If you need to scale a unit of measure, use the scale filter.
Use the string builder processor to add a new attribute (without spaces or special characters) to a log with the result of the provided template. This enables aggregation of different attributes or raw strings into a single attribute.
The template is defined by both raw text and blocks with the syntax %{attribute_path}.
Notes:
This processor only accepts attributes with values or an array of values in the block (see examples in the UI section below.
If an attribute cannot be used (object or array of object), it is replaced by an empty string or the entire operation is skipped depending on your selection.
If a target attribute already exists, it is overwritten by the result of the template.
Results of a template cannot exceed 256 characters.
Define the string builder processor on the Pipelines page:
With the following log, use the template Request %{http.method} %{http.url} was answered with response %{http.status_code} to returns a result. For example:
Request GET https://app.datadoghq.com/users was answered with response 200
Note: http is an object and cannot be used in a block (%{http} fails), whereas %{http.method}, %{http.status_code}, or %{http.url} returns the corresponding value. Blocks can be used on arrays of values or on a specific attribute within an array.
For example, adding the block %{array_ids} returns:
123,456,789
%{array_users} does not return anything because it is a list of objects. However, %{array_users.first_name} returns a list of first_names contained in the array:
The geoIP parser takes an IP address attribute and extracts continent, country, subdivision, or city information (if available) in the target attribute path.
Most elements contain a name and iso_code (or code for continent) attribute. subdivision is the first level of subdivision that the country uses such as “States” for the United States or “Departments” for France.
For example, the geoIP parser extracts location from the network.client.ip attribute and stores it into the network.client.geoip attribute:
{"type":"geo-ip-parser","name":"Parse the geolocation elements from network.client.ip attribute.","is_enabled":true,"sources":["network.client.ip"],"target":"network.client.geoip"}
Parameter
Type
Required
Description
type
String
Yes
Type of the processor.
name
String
No
Name of the processor.
is_enabled
Boolean
No
If the processors is enabled or not. Default: false.
sources
Array of strings
No
Array of source attributes. Default: network.client.ip.
target
String
Yes
Name of the parent attribute that contains all the extracted details from the sources. Default: network.client.geoip.
Use the lookup processor to define a mapping between a log attribute and a human readable value saved in a Reference Table or the processors mapping table.
For example, you can use the lookup processor to map an internal service ID into a human readable service name. Alternatively, you can use it to check if the MAC address that just attempted to connect to the production environment belongs to your list of stolen machines.
The lookup processor performs the following actions:
Looks if the current log contains the source attribute.
Checks if the source attribute value exists in the mapping table.
If it does, creates the target attribute with the corresponding value in the table.
Optionally, if it does not find the value in the mapping table, it creates a target attribute with the default fallback value set in the fallbackValue field. You can manually enter a list of source_key,target_value pairs or upload a CSV file on the Manual Mapping tab.
The size limit for the mapping table is 100Kb. This limit applies across all Lookup Processors on the platform. However, Reference Tables support larger file sizes.
Optionally, if it does not find the value in the mapping table, it creates a target attribute with the value of the reference table. You can select a value for a Reference Table on the Reference Table tab.
If the processor is enabled or not. Default: false.
source
String
Yes
Source attribute used to perform the lookup.
target
String
Yes
Name of the attribute that contains the corresponding value in the mapping list or the default_lookup if not found in the mapping list.
lookup_table
Array of strings
Yes
Mapping table of values for the source attribute and their associated target attribute values, formatted as [ “source_key1,target_value1”, “source_key2,target_value2” ].
default_lookup
String
No
Value to set the target attribute if the source value is not found in the list.
