microcosm.blue
I noticed defaults were hardcoded
./slingshot --jetstream us-east-1 --cache-dir ./foyer \
--cache-memory-mb 128 \
--cache-disk-gb 2
+4
-2
slingshot/src/firehose_cache.rs
+4
-2
slingshot/src/firehose_cache.rs
···
4
5
pub async fn firehose_cache(
6
cache_dir: impl AsRef<Path>,
7
) -> Result<HybridCache<String, CachedRecord>, String> {
8
let cache = HybridCacheBuilder::new()
9
.with_name("firehose")
10
-
.memory(64 * 2_usize.pow(20))
11
.with_weighter(|k: &String, v| k.len() + std::mem::size_of_val(v))
12
.storage(Engine::large())
13
.with_device_options(
14
DirectFsDeviceOptions::new(cache_dir)
15
-
.with_capacity(2_usize.pow(30)) // TODO: configurable (1GB to have something)
16
.with_file_size(16 * 2_usize.pow(20)), // note: this does limit the max cached item size, warning jumbo records
17
)
18
.build()
···
4
5
pub async fn firehose_cache(
6
cache_dir: impl AsRef<Path>,
7
+
memory_mb: usize,
8
+
disk_gb: usize,
9
) -> Result<HybridCache<String, CachedRecord>, String> {
10
let cache = HybridCacheBuilder::new()
11
.with_name("firehose")
12
+
.memory(memory_mb * 2_usize.pow(20))
13
.with_weighter(|k: &String, v| k.len() + std::mem::size_of_val(v))
14
.storage(Engine::large())
15
.with_device_options(
16
DirectFsDeviceOptions::new(cache_dir)
17
+
.with_capacity(disk_gb * 2_usize.pow(30))
18
.with_file_size(16 * 2_usize.pow(20)), // note: this does limit the max cached item size, warning jumbo records
19
)
20
.build()
+22
-5
slingshot/src/main.rs
+22
-5
slingshot/src/main.rs
···
25
/// where to keep disk caches
26
#[arg(long)]
27
cache_dir: PathBuf,
28
/// the domain pointing to this server
29
///
30
/// if present:
···
62
63
let args = Args::parse();
64
65
-
if let Err(e) = install_metrics_server() {
66
log::error!("failed to install metrics server: {e:?}");
67
} else {
68
-
log::info!("metrics listening at http://0.0.0.0:8765");
69
}
70
71
std::fs::create_dir_all(&args.cache_dir).map_err(|e| {
···
83
log::info!("cache dir ready at at {cache_dir:?}.");
84
85
log::info!("setting up firehose cache...");
86
-
let cache = firehose_cache(cache_dir.join("./firehose")).await?;
87
log::info!("firehose cache ready.");
88
89
let mut tasks: tokio::task::JoinSet<Result<(), MainTaskError>> = tokio::task::JoinSet::new();
···
112
args.domain,
113
args.acme_contact,
114
args.certs,
115
server_shutdown,
116
)
117
.await?;
···
172
Ok(())
173
}
174
175
-
fn install_metrics_server() -> Result<(), metrics_exporter_prometheus::BuildError> {
176
log::info!("installing metrics server...");
177
let host = [0, 0, 0, 0];
178
-
let port = 8765;
179
PrometheusBuilder::new()
180
.set_quantiles(&[0.5, 0.9, 0.99, 1.0])?
181
.set_bucket_duration(std::time::Duration::from_secs(300))?
···
25
/// where to keep disk caches
26
#[arg(long)]
27
cache_dir: PathBuf,
28
+
/// memory cache size in MB
29
+
#[arg(long, default_value_t = 64)]
30
+
cache_memory_mb: usize,
31
+
/// disk cache size in GB
32
+
#[arg(long, default_value_t = 1)]
33
+
cache_disk_gb: usize,
34
+
/// port for HTTP server (when not using --domain)
35
+
#[arg(long, default_value_t = 3000)]
36
+
port: u16,
37
+
/// port for metrics/prometheus server
38
+
#[arg(long, default_value_t = 8765)]
39
+
metrics_port: u16,
40
/// the domain pointing to this server
41
///
42
/// if present:
···
74
75
let args = Args::parse();
76
77
+
if let Err(e) = install_metrics_server(args.metrics_port) {
78
log::error!("failed to install metrics server: {e:?}");
79
} else {
80
+
log::info!("metrics listening at http://0.0.0.0:{}", args.metrics_port);
81
}
82
83
std::fs::create_dir_all(&args.cache_dir).map_err(|e| {
···
95
log::info!("cache dir ready at at {cache_dir:?}.");
96
97
log::info!("setting up firehose cache...");
98
+
let cache = firehose_cache(
99
+
cache_dir.join("./firehose"),
100
+
args.cache_memory_mb,
101
+
args.cache_disk_gb,
102
+
)
103
+
.await?;
104
log::info!("firehose cache ready.");
105
106
let mut tasks: tokio::task::JoinSet<Result<(), MainTaskError>> = tokio::task::JoinSet::new();
···
129
args.domain,
130
args.acme_contact,
131
args.certs,
132
+
args.port,
133
server_shutdown,
134
)
135
.await?;
···
190
Ok(())
191
}
192
193
+
fn install_metrics_server(port: u16) -> Result<(), metrics_exporter_prometheus::BuildError> {
194
log::info!("installing metrics server...");
195
let host = [0, 0, 0, 0];
196
PrometheusBuilder::new()
197
.set_quantiles(&[0.5, 0.9, 0.99, 1.0])?
198
.set_bucket_duration(std::time::Duration::from_secs(300))?
+7
-1
slingshot/src/server.rs
+7
-1
slingshot/src/server.rs
···
694
domain: Option<String>,
695
acme_contact: Option<String>,
696
certs: Option<PathBuf>,
697
shutdown: CancellationToken,
698
) -> Result<(), ServerError> {
699
let repo = Arc::new(repo);
···
752
)
753
.await
754
} else {
755
-
run(TcpListener::bind("127.0.0.1:3000"), app, shutdown).await
756
}
757
}
758
···
694
domain: Option<String>,
695
acme_contact: Option<String>,
696
certs: Option<PathBuf>,
697
+
port: u16,
698
shutdown: CancellationToken,
699
) -> Result<(), ServerError> {
700
let repo = Arc::new(repo);
···
753
)
754
.await
755
} else {
756
+
run(
757
+
TcpListener::bind(format!("127.0.0.1:{port}")),
758
+
app,
759
+
shutdown,
760
+
)
761
+
.await
762
}
763
}
764
History
4 rounds
4 comments
nekomimi.pet
submitted
#3
1 commit
expand
collapse
slingshot: add configurable cache sizes, host, and ports
- Add --cache-memory-mb and --cache-disk-gb flags
- Add --host and --port for HTTP server
- Add --metrics-port for prometheus metrics server
expand 4 comments
thank you!!!
the cache size config is perfect, host and port close: i've started moving to a bind arg and giving clap the SocketAddr type directly so it gets parsed out early + one less thing to pass around.
bind stuff came in through a different pr that i just picked up to get merged, so i'm going to get these cache sizes in by picking this up and extracting that for merging
thanks for getting it started!
closed without merging
nekomimi.pet
submitted
#2
1 commit
expand
collapse
slingshot: add configurable cache sizes, host, and ports
- Add --cache-memory-mb and --cache-disk-gb flags
- Add --host and --port for HTTP server
- Add --metrics-port for prometheus metrics server
expand 0 comments
nekomimi.pet
submitted
#1
expand 0 comments
nekomimi.pet
submitted
#0
1 commit
expand
collapse
slingshot: configurable cache sizes
sorry about the messy history, i ordered the git reset soft and amend wrong