Struct IntervalsSection 

pub struct IntervalsSection {
    pub blob_collect_secs: Option<u64>,
    pub confinement_poll_secs: Option<u64>,
    pub health_secs: Option<u64>,
    pub tree_reload_secs: Option<u64>,
}

How often each background loop runs.

Every one of these trades freshness against load, and they are set apart because they load different things: one talks to a DNS provider with its own rate limits, one walks this server’s own disk, and one only reads state already in memory.

Reload: needs a restart. Each loop is handed its period when it is spawned.

Fields

blob_collect_secs: Option<u64>

Seconds between passes of the blob collector, which reclaims upload bytes nothing references once their grace window has passed.

Reasonable range: 300 to 86_400 — five minutes to a day.

This is how promptly deleted bytes come back, and each pass walks every account’s blobs. Five minutes suits a deployment near its quota, where the disk is the scarce thing. A day suits one with room to spare, where the walk is the more expensive half. Nothing is lost by a long period: a blob past its grace window stays collectable until a pass reaches it. The built-in default is an hour.

confinement_poll_secs: Option<u64>

Seconds between passes of the confinement poll, which notices an account whose hostname no longer resolves the way this server published it.

Reasonable range: 60 to 3600 — a minute to an hour.

Each pass reads records back from the DNS provider, so this is the knob that decides how much of a provider’s rate limit this deployment spends on noticing drift. A minute catches a hijacked or clobbered record fast and costs the most requests; an hour is cheaper and leaves a wrong record standing longer. The built-in default is five minutes.

health_secs: Option<u64>

Seconds between health ticks, which refresh what bot.did.stats and the dashboard report about this server’s own state.

Reasonable range: 5 to 300 — five seconds to five minutes.

This one costs nothing outside the process: it reads state already in memory, so the period is how stale a readout may be rather than a load question. Match it to whatever scrapes the dashboard. The built-in default is ten seconds.

tree_reload_secs: Option<u64>

Seconds between re-reads of this configuration file for [tree]’s three bounds, which take effect without a restart.

Reasonable range: 5 to 3600 — five seconds to an hour.

Each pass reads and parses one local file, so the period is how long an operator waits for an edit to take rather than a load question. Five seconds suits raising a cap that is refusing work right now; an hour suits a file nobody edits. The built-in default is thirty seconds.

Trait Implementations

impl Clone for IntervalsSection

fn clone(&self) -> IntervalsSection

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for IntervalsSection

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for IntervalsSection

fn default() -> IntervalsSection

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for IntervalsSection

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for IntervalsSection

fn eq(&self, other: &IntervalsSection) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for IntervalsSection

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl Eq for IntervalsSection

impl StructuralPartialEq for IntervalsSection