blob: 8548655d4cf9ae27532df4f6caf7d0e3c2319455 (
plain) (
blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
|
#pragma once
#include "public.h"
#include <yt/yt/client/table_client/config.h>
#include <yt/yt/core/ytree/yson_struct.h>
#include <yt/yt/core/logging/config.h>
#include <yt/yt/core/misc/backoff_strategy.h>
namespace NYT::NLogging {
////////////////////////////////////////////////////////////////////////////////
//! This log writer can write to both sorted and ordered tables.
//! Configuration notes:
//! - This writer *must* be configured with format = yson.
//! - System events are disabled by default for format = yson logs. If you are enabling them for family = plain_text
//! logs, do not forget to set system_message_family = plain_text.
//! - If multiple hosts are writing to the same table, consider setting enable_host_field = true.
class TDynamicTableLogWriterConfig
: public TLogWriterConfig
{
public:
static constexpr const TStringBuf WriterType = "dynamic_table";
//! Log destination.
NYPath::TYPath TablePath;
//! Log events will be flushed with at least this frequency.
TDuration FlushPeriod;
//! Due to its asynchronous nature, the writer keeps serialized events in its own internal buffer.
//! The buffer size is controlled by the two limits below. Both of them are specified in *bytes*.
//! If the buffer size exceeds the high watermark, the writer drops any new events until the buffer
//! size drops below the low watermark.
i64 LowBacklogWeightWatermark;
i64 HighBacklogWeightWatermark;
//! Controls the maximum number of rows that will be batched within a single transaction.
//! Increase very carefully, the typical server limit is 100'000 rows per transaction.
//! You can decrease this limit to make your transactions smaller.
i64 MaxBatchRowCount;
//! Controls the maximum number of *bytes* that will be batched within a single transaction.
//! You can decrease this limit to make your transactions smaller.
i64 MaxBatchWeight;
//! Infinite retries are configured by default. In this case, persistent write failures will eventually cause
//! the buffer to overflow and the drop of all incoming events.
//! One can configure a small number of invocations, which would allow the logger to progress through persistent write failures
//! by dropping events from the beginning of the queue.
TExponentialBackoffOptions WriteBackoff;
//! YSON-parsing options, similar to the ones available in insert_rows.
//! They are used to convert YSON-serialized log events to unversioned rows.
NTableClient::TTypeConversionConfigPtr TypeConversion;
NTableClient::TInsertRowsFormatConfigPtr InsertRowsFormat;
REGISTER_YSON_STRUCT(TDynamicTableLogWriterConfig);
static void Register(TRegistrar registrar);
};
DEFINE_REFCOUNTED_TYPE(TDynamicTableLogWriterConfig)
////////////////////////////////////////////////////////////////////////////////
} // namespace NYT::NLogging
|
