December 20, 2018: Removed two graphs Total relay bandwidth and Consumed bandwidth by Exit/Guard flag combination, and updated the data format of the Advertised and consumed bandwidth by relay flags graph to cover all data previously contained in the first two graphs. December 20, 2018: Removed source parameters and output rows with aggregates over all sources from Time to download files over Tor, Timeouts and failures of downloading files over Tor, Circuit build times, Circuit round-trip latencies graphs. October 28, 2018: Added and/or removed columns to Tor Browser downloads and updates by platform and Tor Browser downloads and updates by locale graphs. September 15, 2018: Removed all pre-aggregated CSV files. August 15, 2018: Made the first batch of changes to per-graph CSV files. July 31, 2018: Announced pending changes to per-graph CSV files to become effective on August 15 and pre-aggregated CSV files to be removed by September 15. May 29, 2018: Made all parameters of per-graph CSV files optional to support providing both pre-filtered and complete data sets. February 28, 2018: Added per-graph CSV files to eventually replace pre-aggregated CSV files. The following changes have been made in the past or are scheduled to become effective in the near future: ChangesĬhanges to columns will be announced on this page at least a couple weeks in advance as well as on the mailing list. Applications should be able to handle newly added columns and fail gracefully in case of removed columns. Instead, applications should refer to columns by their name. This rule of thumb is not yet implemented for all per-graph statistics files.Īpplications must not rely on the order of columns, as this order may change when columns are removed. The rule of thumb for columns is that neither the choice of parameters nor availability of data should affect the set of columns, but that only a code change can add, change, or remove a column. Columns are pre-defined and specified further down below. ColumnsĮach per-graph statistics file starts with a comment section, followed by a header line and then the actual data lines. In reverse, omitting a parameter produces a larger statistics file that is not filtered by that parameter. Including a parameter in a URL typically filters the resulting statistics file by the given parameter value. Typically, these are start and end plus additional parameters as specified below. These URLs all support a set of optional parameters that can be used to further customize their content. 
ParametersĪll per-graph statistics files are available for download via an URL of the form: This page contains specifications of statistics files used on this website.
0 kommentar(er)
