Get desktop application:
View/edit binary Protocol Buffers messages
all service methods are captilized compared to other regular methods
Sends a greeting
Get all available data sources
Get table state for given table
columns = dimension columns + metric columns metric column are column with number types, others are dimension columns
column histograms serialized in JSON string
Query Nebula to get result
define query request
at the top tree as filter - it can be only one, either AND or OR
every query will specify time range for the data to query - this is enforced start time and end time are unix timestamp with GMT timezone - no delta time range filter will be inclusive on both ends as [start, end]
window size is used in time serials query that aggregate timestamp into a window as bucket this is in seconds (if the value is 60 -> one minute)
dimensions (key columns / group by columns) to pull metrics for
metric list, such as sum(col_a), count(col_b)
usually query only takes top X results based on order spec
define order spec
display type for query result
custom column list
custom time unit
timezone offset in minutes (following javascript getTimezoneOffset convention) timezone behind GMT is positive, ahead is negative (how to bring it back to GMT)
define query response from server
on demand loading specified data with parameters template defined in cluster configuration
on-demand data loading request
load type
template table name
different type will utilize this json object for different purpose. configured template uses it as parameters instantiate a template. google sheets will use it store google sheet spec
TTL in seconds for the data to live
load command response
load state: error=0 indicating successâ
loading time in milliseconds
loaded instance table name
extra message the server wants to share
shut down all work nodes - used for perf profiling
URL service - code for raw or raw for code
Url shorten service
Discovery service - register node
IP address
service port
host name (optional)
service role/tier - server, node, web, etc.
(message has no fields)
define custom column described by a script
Used in:
define supported custom column type which is a sub set of nebula types
Used in:
Used in:
use native flat buffer format it has its own serialziation and deserialization logic
JSON string sending in bytes buffer
The request message containing the user's name.
Used as request type in: V1.Echo, V1.Nuclear
The response message containing the greetings
Used as response type in: V1.Echo, V1.Nuclear
load error
Used in:
Used in:
load command to load data preconfigured (cluster.yml)
A spec to load google sheet
A spec to load a data file(s) by auto detecting its schema
A simple command to unload a table
A metric is defined by rollup method on a column
Used in:
single predicate which is column op [list of values] Operation defines comparions op to form a predicate: =, !=, >, <
Used in:
Used in:
define ordering spec - limiting to order by single column in returning result
Used in:
TODO(cao): reference Constraints.js for GROUP/RULE generation to build predicates tree Current definition only supports one group basically. (nested groups) NOTE - based on the code generation rule, we don't put plural term for repeated field so the method will like add_value rather than add_values Based on column type - the value will be converted from string always Adding n_value for integers and d_value for floating values if client uses it. Adding values as customized data list serialzied in some format (data type same as column type)
Used in:
,allow client to pass a zipped values representing an list of values in current column type. zip format and zip count show what algo used to zip the original array. zip count indicates how many values zipped inside.
define filter - a filter is formed by a list of predicates combined by AND or OR
Used in:
Used in:
This defines all the aggregate metric methods suppored in nebula
Used in:
pre-defined percentile
call stack merge
estimated cardinality
Used in:
define query processing metrics
Used in:
end to end query time in milliseconds
total rows scanned for the query
total blocks scanned for the query
total number of rows in result (before limit)
query result - 0=success, other error code to look up
may place error message here if failed
Used in: